Diff of /sml/trunk/HISTORY

Parent Directory | Revision Log | Patch

	revision 658, Mon Jun 12 02:39:37 2000 UTC	revision 1211, Tue May 21 14:54:32 2002 UTC
#	Line 8	Line 8
8	The form of an entry should be:	The form of an entry should be:
9
10	Name:	Name:
11	Date:	Date: yyyy/mm/dd
12	Tag: <post-commit CVS tag>	Tag: <post-commit CVS tag>
13	Description:	Description:
14
15		----------------------------------------------------------------------
16		Name: Matthias Blume
17		Date: 2002/05/21 10:55:00 EDT
18		Tag: blume-20020521-misc
19		Description:
20
21		- John tweaked runtime to be silent on heap export (except when
22		GC messages are on).
23		- I added a few more things (cross-compiling versions of CMB) to
24		config/preloads (as suggestions).
25
26		----------------------------------------------------------------------
27		Name: Matthias Blume
28		Date: 2002/05/20 22:25:00 EDT
29		Tag: blume-20020520-controls
30		Description:
31
32		- Added ControlUtil structure to control-lib.cm.
33		- Use it throughout.
34		- Used Controls facility to define MLRISC controls (as opposed to
35		registering MLRISC control ref cells with Controls after the fact)
36		- Fixed messed-up controls priorities.
37
38		* Removed again all the stuff from config/preloads that one wouldn't
39		be able to preload at the time the initial heap image is built.
40		(Many libraries, e.g., CML, do not exist yet at this time.  The
41		only libraries that can be preloaded via config/preloads are those
42		that come bundled with the bootfiles.)
43
44		----------------------------------------------------------------------
45		Name: Matthias Blume
46		Date: 2002/05/20 10:59:00 EDT
47		Tag: blume-20020520-preloads
48		Description:
49
50		Added a lot of commented-out suggestions for things to be included
51		in config/preloads.
52
53		----------------------------------------------------------------------
54		Name: Allen Leung
55		Date: 2002/05/18 14:20:00 EDT
56		Tag: leunga-20020518-mdl
57		Description:
58
59		o Made the mdl tool stuff compile and run again.
60		o I've disabled all the stuff that depends on RTL specifications; they
61		are all badly broken anyway.
62
63		----------------------------------------------------------------------
64		Name: Matthias Blume
65		Date: 2002/05/17 16:49:00 EDT
66		Tag: blume-20020517-controls
67		Description:
68
69		0. John Reppy made several modifications to the SML/NJ library.
70		In particular, there is a shiny new controls-lib.cm.
71
72		1. Pushed new controls interface through compiler so that everything
73		compiles again.
74
75		2. Added FormatComb and FORMAT_COMB to the CML version of the
76		SML/NJ library (so that CML compiles again).
77
78		3. Modified init scripts because XXX_DEFAULT environment variables
79		are no longer with us.  (Boot-time initialization is now done
80		using the same environment variables that are also used for
81		startup-time initialization of controls.)
82
83		----------------------------------------------------------------------
84		Name: Lal George
85		Date: 2002/05/15 09:20:10 EDT
86		Tag: george-20020515-pseudo-op-decls
87		Description:
88
89		All pseudo-ops emitted before the first segment declaration
90		such as TEXT, DATA, and BSS directives are assumed to be global
91		declarations and are emitted first in the assembly file. This is
92		useful in a number of situations where one has pseudo-ops that are not
93		specific to any segment, and also works around the constraint that one
94		cannot have client pseudo-ops in the TEXT segment.
95
96		Because no segment is associated with these declarations it is
97		an error to allocate any space or objects before the first segment
98		directive and an exception will be raised. However, we cannot make
99		this check for client pseudo-ops.
100
101		These top level declarations are a field in the CFG graph_info.
102		In theory you can continue to add to this field after the CFG has been
103		built -- provided you know what you are doing;-)
104
105		----------------------------------------------------------------------
106		Name: Matthias Blume
107		Date: 2002/05/13 16:40:00 EDT
108		Tag: blume-20020513-pp-etc
109		Description:
110
111		A few minor bugfixes:
112
113		- Stopgap measure for bug recently reported by Elsa Gunter (ppDec).
114		(Bogus printouts for redefined bindings still occur.  Compiler
115		bug should no longer occur now.  We need to redo the prettyprinter
116		from scratch.)
117
118		- CM pathname printer now also adds escape sequences for ( and )
119
120		- commend and docu fixes for ml-nlffi
121
122		----------------------------------------------------------------------
123		Name: Matthias Blume
124		Date: 2002/05/10 16:40:00 EDT
125		Tag: blume-20020510-erg-textio
126		Description:
127
128		Applied the following bugfix provided by Emden Gansner:
129
130		Output is corrupted when outputSubstr is used rather than output.
131
132		The problem occurs when a substring
133
134		ss = (s, dataStart, dataLen)
135
136		where dataStart > 0, fills a stream buffer with avail bytes left.
137		avail bytes of s, starting at index dataStart, are copied into the
138		buffer, the buffer is flushed, and then the remaining dataLen-avail
139		bytes of ss are copied into the beginning of the buffer. Instead of
140		starting this copy at index dataStart+avail in s, the current code
141		starts the copy at index avail.
142
143		Fix:
144		In text-io-fn.sml, change line 695 from
145		val needsFlush = copyVec(v, avail, dataLen-avail, buf, 0)
146		to
147		val needsFlush = copyVec(v, dataStart+avail, dataLen-avail, buf, 0)
148
149		----------------------------------------------------------------------
150		Name: Matthias Blume
151		Date: 2002/04/12 13:55:00 EDT
152		Tag: blume-20020412-assyntax
153		Description:
154
155		1. Grabbed newer assyntax.h from the XFree86 project.
156		2. Fiddled with how to compile X86.prim.asm without warnings.
157		3. (Very) Minor cleanup in CM.
158
159		----------------------------------------------------------------------
160		Name: Matthias Blume
161		Date: 2002/04/01 (no joke!) 17:07:00 EST
162		Tag: blume-20020401-x86div
163		Description:
164
165		Added full support for div/mod/rem/quot on the x86, using the machine
166		instruction's two results (without clumsily recomputing the remainder)
167		directly where appropriate.
168
169		Some more extensive power-of-two support was added to the x86 instruction
170		selector (avoiding expensive divs, mods, and muls where they can be
171		replaced with cheaper shifts and masks).  However, this sort of thing
172		ought to be done earlier, e.g., within the CPS optimizer so that
173		all architectures benefit from it.
174
175		The compiler compiles to a fixed point, but changes might be somewhat
176		fragile nevertheless.  Please, report any strange things that you might
177		see wrt. div/mod/quot/rem...
178
179		----------------------------------------------------------------------
180		Name: Matthias Blume
181		Date: 2002/03/29 17:22:00
182		Tag: blume-20020329-div
183		Description:
184
185		Fixed my broken div/mod logic.  Unfortunately, this means that the
186		inline code for div/mod now has one more comparison than before.
187		Fast paths (quotient > 0 or remainder = 0) are not affected, though.
188		The problem was with quotient = 0, because that alone does not tell
189		us which way the rounding went.  One then has to look at whether
190		remainder and divisor have the same sign...  :(
191
192		Anyway, I replaced the bootfiles with fresh ones...
193
194		----------------------------------------------------------------------
195		Name: Matthias Blume
196		Date: 2002/03/29 14:10:00 EST
197		Tag: blume-20020329-inlprims
198		Description:
199
200		NEW BOOTFILES!!!    Version number bumped to 110.39.3.
201
202		Primops have changed. This means that the bin/boot-file formats have
203		changed as well.
204
205		To make sure that there is no confusion, I made a new version.
206
207
208		CHANGES:
209
210		* removed REMT from mltree (remainder should never overflow).
211
212		* added primops to deal with divisions of all flavors to the frontend
213
214		* handled these primops all the way through so they map to their respective
215		MLRISC support
216
217		* used these primops in the implementation of Int, Int32, Word, Word32
218
219		* removed INLDIV, INLMOD, and INLREM as they are no longer necessary
220
221		* parameterized INLMIN, INLMAX, and INLABS by a numkind
222
223		* translate.sml now deals with all flavors of INL{MIN,MAX,ABS}, including
224		floating point
225
226		* used INL{MIN,MAX,ABS} in the implementation of Int, Int32, Word, Word32,
227		and Real (but Real.abs maps to a separate floating-point-only primop)
228
229
230		TODO items:
231
232		* Hacked Alpha32 instruction selection, disabling the selection of REMx
233		instructions because the machine instruction encoder cannot handle
234		them.  (Hppa, PPC, and Sparc instruction selection did not handle
235		REM in the first place, and REM is supported by the x86 machine coder.)
236
237		* Handle DIV and MOD with DIV_TO_NEGINF directly in the x86 instruction
238		selection phase.  (The two can be streamlined because the hardware
239		delivers both quotient and remainder at the same time anyway.)
240
241		* Think about what to do with "valOf(Int32.minInt) div ~1" and friends.
242		(Currently the behavior is inconsistent both across architectures and
243		wrt. the draft Basis spec.)
244
245		* Word8 should eventually be handled natively, too.
246
247		* There seems to be one serious bug in mltree-gen.sml.  It appears, though,
248		as if there currently is no execution path that could trigger it in
249		SML/NJ.  (The assumptions underlying functions arith and promotable do not
250		hold for things like multiplication and division.)
251
252		----------------------------------------------------------------------
253		Name: Matthias Blume
254		Date: 2002/03/27 16:27:00 EST
255		Tag: blume-20020327-mlrisc-divisions
256		Description:
257
258		Added support for all four division operations (ML's div, mod, quot,
259		and rem) to MLRISC.  In the course of doing so, I also rationalized
260		the naming (no more annoying switch-around of DIV and QUOT), by
261		parameterizing the operation by div_rounding_mode (which can be either
262		DIV_TO_ZERO or DIV_TO_NEGINF).
263
264		The generic MLTreeGen functor takes care of compiling all four
265		operations down to only round-to-zero div.
266
267		Missing pieces:
268
269		* Doing something smarter than relying on MLTreeGen on architectures
270		like, e.g., the x86 where hardware division delivers both quotient and
271		remainder at the same time.  With this, the implementation of the
272		round-to-neginf operations could be further streamlined.
273
274		* Remove inlining support for div/mod/rem from the frontend and replace it
275		with primops that get carried through to the backend.  Do this for all
276		int and word types.
277
278		----------------------------------------------------------------------
279		Name: Matthias Blume
280		Date: 2002/03/25 17:25:00 EST
281		Tag: blume-20020325-divmod
282		Description:
283
284		I improved (hopefully without breaking them) the implementation of Int.div,
285		Int.mod, and Int.rem.   For this, the code in translate.sml now takes
286		advantage of the following observations:
287
288		Let  q = x quot y      r = x rem y
289		d = x div  y      m = x mod y
290
291		where "quot" is the round-to-zero version of integer division that
292		hardware usually provides.  Then we have:
293
294		r = x - q * y        where neither the * nor the - will overflow
295		d = if q >= 0 orelse x = q * y then q else q - 1
296		where neither the * nor the - will overflow
297		m = if q >= 0 orelse r = 0 then r else r + y
298		where the + will not overflow
299
300		This results in substantial simplification of the generated code.
301		The following table shows the number of CFG nodes and edges generated
302		for
303		fun f (x, y) = x OPER y
304		(* with OPER \in div, mod, quot, rem *)
305
306
307		OPER | nodes(old) | edges(old) | nodes(new) | edges(new)
308		--------------------------------------------------------
309		div |         24 |         39 |         12 |         16
310		mod |         41 |         71 |         12 |         16
311		quot |          8 |         10 |          8 |         10
312		rem |         10 |         14 |          8 |         10
313
314
315		----------------------------------------------------------------------
316		Name: Matthias Blume
317		Date: 2002/03/25 22:06:00 EST
318		Tag: blume-20020325-cprotobug
319		Description:
320
321		Fixed a bug in cproto (c prototype decoder).
322
323		----------------------------------------------------------------------
324		Name: Matthias Blume
325		Date: 2002/03/25 16:00:00 EST
326		Tag: blume-20020325-raw-primops
327		Description:
328
329		I did some cleanup to Allen's new primop code and
330		replaced yesterday's bootfiles with new ones.
331		(But they are stored in the same place.)
332
333		----------------------------------------------------------------------
334		Name: Matthias Blume
335		Date: 2002/03/24 22:40:00 EST
336		Tag: blume-20020324-bootfiles
337		Description:
338
339		Made the bootfiles that Allen asked for.
340
341		----------------------------------------------------------------------
342		Name: Allen Leung
343		Date: 2002/03/23 15:50:00 EST
344		Tag: leunga-20020323-flint-cps-rcc-primops
345		Description:
346
347		1. Changes to FLINT primops:
348
349		(* make a call to a C-function;
350		* The primop carries C function prototype information and specifies
351		* which of its (ML-) arguments are floating point. C prototype
352		* information is for use by the backend, ML information is for
353		* use by the CPS converter. *)
354		| RAW_CCALL of { c_proto: CTypes.c_proto,
355		ml_args: ccall_type list,
356		ml_res_opt: ccall_type option,
357		reentrant : bool
358		} option
359		(* Allocate uninitialized storage on the heap.
360		* The record is meant to hold short-lived C objects, i.e., they
361		* are not ML pointers.  With the tag, the representation is
362		* the same as RECORD with tag tag_raw32 (sz=4), or tag_fblock (sz=8)
363		*)
364		| RAW_RECORD of {tag:bool,sz:int}
365		and ccall_type = CCALL_INT32 | CCALL_REAL64 | CCALL_ML_PTR
366
367		2.  These CPS primops are now overloaded:
368
369		rawload of {kind:numkind}
370		rawstore of {kind:numkind}
371
372		The one argument form is:
373
374		rawload {kind} address
375
376		The two argument form is:
377
378		rawload {kind} [ml object, byte-offset]
379
380		3. RAW_CCALL/RCC now takes two extra arguments:
381
382		a. The first is whether the C call is reentrant, i.e., whether
383		ML state should be saved and restored.
384		b. The second argument is a string argument specifying the name of
385		library and the C function.
386
387		These things are currently not handled in the code generator, yet.
388
389		4. In CProto,
390
391		An encoding type of "bool" means "ml object" and is mapped into
392		C prototype of PTR.  Note that "bool" is different than "string",
393		even though "string" is also mapped into PTR, because "bool"
394		is assigned an CPS type of BOGt, while "string" is assigned INT32t.
395
396		5. Pickler/unpicker
397
398		Changed to handle RAW_RECORD and newest RAW_CCALL
399
400		6. MLRiscGen,
401
402		1. Changed to handle the new rawload/rawstore/rawrecord operators.
403		2. Code for handling C Calls has been moved to a new module CPSCCalls,
404		in the file CodeGen/cpscompile/cps-c-calls.sml
405
406		7. Added the conditional move operator
407
408		condmove of branch
409
410		to cps.  Generation of this is still buggy so it is currently
411		disabled.
412
413		----------------------------------------------------------------------
414		Name: Lal George
415		Date: 2002/03/22 14:18:25 EST
416		Tag: george-20020322-cps-branch-prob
417		Description:
418
419		Implemented the Ball-Larus branch prediction-heuristics, and
420		incorporated graphical viewers for control flow graphs.
421
422		Ball-Larus Heuristics:
423		---------------------
424		See the file compiler/CodeGen/cpscompile/cpsBranchProb.sml.
425
426		By design it uses the Dempster-Shafer theory for combining
427		probabilities.  For example, in the function:
428
429		fun f(n,acc) = if n = 0 then acc else f(n-1, n*acc)
430
431		the ball-larus heuristics predicts that the n=0 is unlikely
432		(OH-heuristic), and the 'then' branch is unlikely because of the
433		RH-heuristic -- giving the 'then' branch an even lower combined
434		probability using the Dempster-Shafer theory.
435
436		Finally, John Reppy's loop analysis in MLRISC, further lowers the
437		probability of the 'then' branch because of the loop in the else
438		branch.
439
440
441		Graphical Viewing:
442		------------------
443		I merely plugged in Allen's graphical viewers into the compiler. The
444		additional code is not much. At the top level, saying:
445
446		Control.MLRISC.getFlag "cfg-graphical-view" := true;
447
448		will display the graphical view of the control flow graph just before
449		back-patching.  daVinci must be in your path for this to work. If
450		daVinci is not available, then the default viewer can be changed
451		using:
452
453		Control.MLRISC.getString "viewer"
454
455		which can be set to "dot" or "vcg" for the corresponding viewers. Of
456		course, these viewers must be in your path.
457
458		The above will display the compilation unit at the level of clusters,
459		many of which are small, boring, and un-interesting. Also setting:
460
461		Control.MLRISC.getInt "cfg-graphical-view_size"
462
463		will display clusters that are larger than the value set by the above.
464
465
466		----------------------------------------------------------------------
467		Name: Matthias Blume
468		Date: 2002/03/21 22:20:00 EST
469		Tag: blume-20020321-kmp-bugfix
470		Description:
471
472		Changed the interface to the KMP routine in PreString and fixed
473		a minor bug in one place where it was used.
474
475		----------------------------------------------------------------------
476		Name: Allen Leung
477		Date: 2002/03/21 20:30:00 EST
478		Tag: leunga-20020321-cfg
479		Description:
480
481		Fixed a potential problem in cfg edge splitting.
482
483		----------------------------------------------------------------------
484		Name: Allen Leung
485		Date: 2002/03/21 17:15:00 EST
486		Tag: leunga-20020321-x86-fp-cfg
487		Description:
488
489		1. Recoded the buggy parts of x86-fp.
490
491		a. All the block reordering code has been removed.
492		We now depend on the block placement phases to do this work.
493
494		b. Critical edge splitting code has been simplified and moved into the
495		CFG modules, as where they belong.
496
497		Both of these were quite buggy and complex.  The code is now much, much
498		simpler.
499
500		2. X86 backend.
501
502		a. Added instructions for 64-bit support.  Instruction selection for
503		64-bit has not been committed, however, since that
504		requires changes to MLTREE which haven't been approved by
505		Lal and John.
506
507		b. Added support for FUCOMI and FUCOMIP when generating code for
508		PentiumPro and above.  We only generate these instructions in
509		the fast-fp mode.
510
511		c. Added cases for JP and JNP in X86FreqProps.
512
513		3. CFG
514
515		CFG now has a bunch of methods for edge splitting and merging.
516
517		4. Machine description.
518
519		John's simplification of MLTREE_BASIS.fcond broke a few machine
520		description things:
521
522		rtl-build.{sig,sml} and hppa.mdl fixed.
523
524		NOTE: the machine description stuff in the repository is still broken.
525		Again, I can't put my fixes in because that involves
526		changes to MLTREE.
527
528		----------------------------------------------------------------------
529		Name: Matthias Blume
530		Date: 2002/03/20 15:55:00 EST
531		Tag: blume-20020320-kmp
532		Description:
533
534		Implemented Knuth-Morris-Pratt string matching in PreString and used
535		it for String.isSubstring, Substring.isSubstring, and
536		Substring.position.
537
538		(Might need some stress-testing.  Simple examples worked fine.)
539
540		----------------------------------------------------------------------
541		Name: Matthias Blume
542		Date: 2002/03/19 16:37:00 EST
543		Tag: blume-20020319-witnesses
544		Description:
545
546		Added a structure C.W and functions convert/Ptr.convert to ml-nlffi-lib.
547
548		This implements a generic mechanism for changing constness qualifiers
549		anywhere within big C types without resorting to outright "casts".
550		(So far, functions such as C.rw/C.ro or C.Ptr.rw/C.Ptr.ro only let you
551		modify the constness at the outermost level.)
552		The implementation of "convert" is based on the idea of "witness"
553		values -- values that are not used by the operation but whose types
554		"testify" to their applicability.  On the implementation side, "convert"
555		is simply a projection (returning its second curried argument).  With
556		cross-module inlining, it should not result in any machine code being
557		generated.
558
559		----------------------------------------------------------------------
560		Name: Matthias Blume
561		Date: 2002/03/15 16:40:00 EST
562		Tag: blume-20020315-basis
563		Description:
564
565		Provided (preliminary?) implementations for
566
567		{String,Substring}.{concatWith,isSuffix,isSubstring}
568
569		and
570
571		Substring.full
572
573		Those are in the Basis spec but they were missing in SML/NJ.
574
575		----------------------------------------------------------------------
576		Name: Matthias Blume
577		Date: 2002/03/14 21:30:00 EST
578		Tag: blume-20020314-controls
579		Description:
580
581		Controls:
582		---------
583
584		1. Factored out the recently-added Controls : CONTROLS stuff and put
585		it into its own library $/controls-lib.cm.  The source tree for
586		this is under src/smlnj-lib/Controls.
587
588		2. Changed the names of types and functions in this interface, so they
589		make a bit more "sense":
590
591		module -> registry
592		'a registry -> 'a group
593
594		3. The interface now deals in ref cells only.  The getter/setter interface
595		is (mostly) gone.
596
597		4. Added a function that lets one register an already-existing ref cell.
598
599		5. Made the corresponding modifications to the rest of the code so that
600		everything compiles again.
601
602		6. Changed the implementation of Controls.MLRISC back to something closer
603		to the original.  In particular, this module (and therefore MLRISC)
604		does not depend on Controls.  There now is some link-time code in
605		int-sys.sml that registers the MLRISC controls with the Controls
606		module.
607
608		CM:
609		---
610
611		* One can now specify the lambda-split aggressiveness in init.cmi.
612
613		----------------------------------------------------------------------
614		Name: Allen Leung
615		Date: 2002/03/13 17:30:00 EST
616		Tag: leunga-20020313-x86-fp-unary
617		Description:
618
619		Bug fix for:
620
621		> leunga@weaselbane:~/Yale/tmp/sml-dist{21} bin/sml
622		> Standard ML of New Jersey v110.39.1 [FLINT v1.5], March 08, 2002
623		> - fun f(x,(y,z)) = Real.~ y;
624		> [autoloading]
625		> [autoloading done]
626		>       fchsl   (%eax), 184(%esp)
627		> Error: MLRisc bug: X86MCEmitter.emitInstr
628		>
629		> uncaught exception Error
630		>   raised at: ../MLRISC/control/mlriscErrormsg.sml:16.14-16.19
631
632		The problem was that the code generator did not generate any fp registers
633		in this case, and the ra didn't know that it needed to run the X86FP phase to
634		translate the pseudo fp instruction.   This only happened with unary fp
635		operators in certain situations.
636
637		----------------------------------------------------------------------
638		Name: Matthias Blume
639		Date: 2002/03/13 14:00:00 EST
640		Tag: blume-20020313-overload-etc
641		Description:
642
643		1. Added _overload as a synonym for overload for backward compatibility.
644		(Control.overloadKW must be true for either version to be accepted.)
645
646		2. Fixed bug in install script that caused more things to be installed
647		than what was requested in config/targets.
648
649		3. Made CM aware of the (_)overload construct so that autoloading
650		works.
651
652		----------------------------------------------------------------------
653		Name: Matthias Blume
654		Date: 2002/03/12 22:03:00 EST
655		Tag: blume-20020312-url
656		Description:
657
658		Forgot to update BOOT and srcarchiveurl.
659
660		----------------------------------------------------------------------
661		Name: Matthias Blume
662		Date: 2002/03/12 17:30:00 EST
663		Tag: blume-20020312-version110392
664		Description:
665
666		Yet another version number bump (because of small changes to the
667		binfile format).  Version number is now 110.39.2.  NEW BOOTFILES!
668
669		Changes:
670
671		The new pid generation scheme described a few weeks ago was overly
672		complicated.  I implemented a new mechanism that is simpler and
673		provides a bit more "stability":  Once CM has seen a compilation
674		unit, it keeps its identity constant (as long as you do not delete
675		those crucial CM/GUID/* files).  This means that when you change
676		an interface, compile, then go back to the old interface, and
677		compile again, you arrive at the original pid.
678
679		There now also is a mechanism that instructs CM to use the plain
680		environment hash as a module's pid (effectively making its GUID
681		the empty string).  For this, "noguid" must be specified as an
682		option to the .sml file in question within its .cm file.
683		This is most useful for code that is being generated by tools such
684		as ml-nlffigen (because during development programmers tend to
685		erase the tool's entire output directory tree including CM's cached
686		GUIDs).  "noguid" is somewhat dangerous (since it can be used to locally
687		revert to the old, broken behavior of SML/NJ, but in specific cases
688		where there is no danger of interface confusion, its use is ok
689		(I think).
690
691		ml-nlffigen by default generates "noguid" annotations.  They can be
692		turned off by specifying -guid in its command line.
693
694		----------------------------------------------------------------------
695		Name: Lal George
696		Date: 2002/03/12 12 14:42:36 EST
697		Tag: george-20020312-frequency-computation
698		Description:
699
700		Integrated jump chaining and static block frequency into the
701		compiler. More details and numbers later.
702
703		----------------------------------------------------------------------
704		Name: Lal George
705		Date: 2002/03/11 11 22:38:53 EST
706		Tag: george-20020311-jump-chain-elim
707		Description:
708
709		Tested the jump chain elimination on all architectures (except the
710		hppa).  This is on by default right now and is profitable for the
711		alpha and x86, however, it may not be profitable for the sparc and ppc
712		when compiling the compiler.
713
714		The gc test will typically jump to a label at the end of the cluster,
715		where there is another jump to an external cluster containing the actual
716		code to invoke gc. This is to allow factoring of common gc invocation
717		sequences. That is to say, we generate:
718
719		f:
720		testgc
721		ja   L1      % jump if above to L1
722
723		L1:
724		jmp L2
725
726
727		After jump chain elimination the 'ja L1' instructions is converted to
728		'ja L2'. On the sparc and ppc, many of the 'ja L2' instructions may end
729		up being implemented in their long form (if L2 is far away) using:
730
731		jbe     L3      % jump if below or equal to L3
732		jmp     L2
733		L3:
734		...
735
736
737		For large compilation units L2  may be far away.
738
739
740		----------------------------------------------------------------------
741		Name: Matthias Blume
742		Date: 2002/03/11 13:30:00 EST
743		Tag: blume-20020311-mltreeeval
744		Description:
745
746		A functor parameter was missing.
747
748		----------------------------------------------------------------------
749		Name: Allen Leung
750		Date: 2002/03/11 10:30:00 EST
751		Tag: leunga-20020311-runtime-string0
752		Description:
753
754		The representation of the empty string now points to a
755		legal null terminated C string instead of unit.  It is now possible
756		to convert an ML string into C string with InlineT.CharVector.getData.
757		This compiles into one single machine instruction.
758
759		----------------------------------------------------------------------
760		Name: Allen Leung
761		Date: 2002/03/10 23:55:00 EST
762		Tag: leunga-20020310-x86-call
763		Description:
764
765		Added machine generation for CALL instruction (relative displacement mode)
766
767		----------------------------------------------------------------------
768		Name: Matthias Blume
769		Date: 2002/03/08 16:05:00
770		Tag: blume-20020308-entrypoints
771		Description:
772
773		Version number bumped to 110.39.1.  NEW BOOTFILES!
774
775		Entrypoints: non-zero offset into a code object where execution should begin.
776
777		- Added the notion of an entrypoint to CodeObj.
778		- Added reading/writing of entrypoint info to Binfile.
779		- Made runtime system bootloader aware of entrypoints.
780		- Use the address of the label of the first function given to mlriscGen
781		as the entrypoint.  This address is currently always 0, but it will
782		not be 0 once we turn on block placement.
783		- Removed the linkage cluster code (which was The Other Way(tm) of dealing
784		with entry points) from mlriscGen.
785
786		----------------------------------------------------------------------
787		Name: Allen Leung
788		Date: 2002/03/07 20:45:00 EST
789		Tag: leunga-20020307-x86-cmov
790		Description:
791
792		Bug fixes for CMOVcc on x86.
793
794		1. Added machine code generation for CMOVcc
795		2. CMOVcc is now generated in preference over SETcc on PentiumPro or above.
796		3. CMOVcc cannot have an immediate operand as argument.
797
798		----------------------------------------------------------------------
799		Name: Matthias Blume
800		Date: 2002/03/07 16:15:00 EST
801		Tag: blume-20020307-controls
802		Description:
803
804		This is a very large but mostly boring patch which makes (almost)
805		every tuneable compiler knob (i.e., pretty much everything under
806		Control.* plus a few other things) configurable via both the command
807		line and environment variables in the style CM did its configuration
808		until now.
809
810		Try starting sml with '-h' (or, if you are brave, '-H')
811
812		To this end, I added a structure Controls : CONTROLS to smlnj-lib.cm which
813		implements the underlying generic mechanism.
814
815		The interface to some of the existing such facilities has changed somewhat.
816		For example, the MLRiscControl module now provides mkFoo instead of getFoo.
817		(The getFoo interface is still there for backward-compatibility, but its
818		use is deprecated.)
819
820		The ml-build script passes -Cxxx=yyy command-line arguments through so
821		that one can now twiddle the compiler settings when using this "batch"
822		compiler.
823
824		TODO items:
825
826		We should go through and throw out all controls that are no longer
827		connected to anything.  Moreover, we should go through and provide
828		meaningful (and correct!) documentation strings for those controls
829		that still are connected.
830
831		Currently, multiple calls to Controls.new are accepted (only the first
832		has any effect).  Eventually we should make sure that every control
833		is being made (via Controls.new) exactly once.  Future access can then
834		be done using Controls.acc.
835
836		Finally, it would probably be a good idea to use the getter-setter
837		interface to controls rather than ref cells.  For the time being, both
838		styles are provided by the Controls module, but getter-setter pairs are
839		better if thread-safety is of any concern because they can be wrapped.
840
841		*****************************************
842
843		One bug fix: The function blockPlacement in three of the MLRISC
844		backpatch files used to be hard-wired to one of two possibilities at
845		link time (according to the value of the placementFlag).  But (I
846		think) it should rather sense the flag every time.
847
848		*****************************************
849
850		Other assorted changes (by other people who did not supply a HISTORY entry):
851
852		1. the cross-module inliner now works much better (Monnier)
853		2. representation of weights, frequencies, and probabilities in MLRISC
854		changed in preparation of using those for weighted block placement
855		(Reppy, George)
856
857		----------------------------------------------------------------------
858		Name: Lal George
859		Date: 2002/03/07 14:44:24 EST 2002
860		Tag: george-20020307-weighted-block-placement
861
862		Tested the weighted block placement optimization on all architectures
863		(except the hppa) using AMPL to generate the block and edge frequencies.
864		Changes were required in the machine properties to correctly
865		categorize trap instructions. There is an MLRISC flag
866		"weighted-block-placement" that can be used to enable weighted block
867		placement, but this will be ineffective without block/edge
868		frequencies (coming soon).
869
870
871		----------------------------------------------------------------------
872		Name: Lal George
873		Date: 2002/03/05 17:24:48 EST
874		Tag: george-20020305-linkage-cluster
875
876		In order to support the block placement optimization, a new cluster
877		is generated as the very first cluster (called the linkage cluster).
878		It contains a single jump to the 'real' entry point for the compilation
879		unit. Block placement has no effect on the linkage cluster itself, but
880		all the other clusters  have full freedom in the manner in which they
881		reorder blocks or functions.
882
883		On the x86 the typical linkage code that is generated is:
884		----------------------
885		.align 2
886		L0:
887		addl    $L1-L0, 72(%esp)
888		jmp     L1
889
890
891		.align  2
892		L1:
893		----------------------
894
895		72(%esp) is the memory location for the stdlink register. This
896		must contain the address of the CPS function being called. In the
897		above example, it contains the address of  L0; before
898		calling L1 (the real entry point for the compilation unit), it
899		must contain the address for L1, and hence
900
901		addl $L1-L0, 72(%esp)
902
903		I have tested this on all architectures except the hppa.The increase
904		in code size is of course negligible
905
906		----------------------------------------------------------------------
907		Name: Allen Leung
908		Date: 2002/03/03 13:20:00 EST
909		Tag: leunga-20020303-mlrisc-tools
910
911		Added #[ ... ] expressions to mlrisc tools
912
913		----------------------------------------------------------------------
914		Name: Matthias Blume
915		Date: 2002/02/27 12:29:00 EST
916		Tag: blume-20020227-cdebug
917		Description:
918
919		- made types in structure C and C_Debug to be equal
920		- got rid of code duplication (c-int.sml vs. c-int-debug.sml)
921		- there no longer is a C_Int_Debug (C_Debug is directly derived from C)
922
923		----------------------------------------------------------------------
924		Name: Matthias Blume
925		Date: 2002/02/26 12:00:00 EST
926		Tag: blume-20020226-ffi
927		Description:
928
929		1. Fixed a minor bug in CM's "noweb" tool:
930		If numbering is turned off, then truly don't number (i.e., do not
931		supply the -L option to noweb).  The previous behavior was to supply
932		-L'' -- which caused noweb to use the "default" line numbering scheme.
933		Thanks to Chris Richards for pointing this out (and supplying the fix).
934
935		2. Once again, I reworked some aspects of the FFI:
936
937		A. The incomplete/complete type business:
938
939		- Signatures POINTER_TO_INCOMPLETE_TYPE and accompanying functors are
940		gone!
941		- ML types representing an incomplete type are now *equal* to
942		ML types representing their corresponding complete types (just like
943		in C).  This is still safe because ml-nlffigen will not generate
944		RTTI for incomplete types, nor will it generate functions that
945		require access to such RTTI.   But when ML code generated from both
946		incomplete and complete versions of the C type meet, the ML types
947		are trivially interoperable.
948
949		NOTE:  These changes restore the full generality of the translation
950		(which was previously lost when I eliminated functorization)!
951
952		B. Enum types:
953
954		- Structure C now has a type constructor "enum" that is similar to
955		how the "su" constructor works.  However, "enum" is not a phantom
956		type because each "T enum" has values (and is isomorphic to
957		MLRep.Signed.int).
958		- There are generic access operations for enum objects (using
959		MLRep.Signed.int).
960		- ml-nlffigen will generate a structure E_foo for each "enum foo".
961		* The structure contains the definition of type "mlrep" (the ML-side
962		representation type of the enum).  Normally, mlrep is the same
963		as "MLRep.Signed.int", but if ml-nlffigen was invoked with "-ec",
964		then mlrep will be defined as a datatype -- thus facilitating
965		pattern matching on mlrep values.
966		("-ec" will be suppressed if there are duplicate values in an
967		enumeration.)
968		* Constructors ("-ec") or values (no "-ec") e_xxx of type mlrep
969		will be generated for each C enum constant xxx.
970		* Conversion functions m2i and i2m convert between mlrep and
971		MLRep.Signed.int.  (Without "-ec", these functions are identities.)
972		* Coversion functions c and ml convert between mlrep and "tag enum".
973		* Access functions (get/set) fetch and store mlrep values.
974		- By default (unless ml-nlffigen was invoked with "-nocollect"), unnamed
975		enumerations are merged into one single enumeration represented by
976		structure E_'.
977
978		----------------------------------------------------------------------
979		Name: Allen Leung
980		Date: 2002/02/25 04:45:00 EST
981		Tag: leunga-20020225-cps-spill
982
983		This is a new implementation of the CPS spill phase.
984		The new phase is in the new file compiler/CodeGen/cpscompile/spill-new.sml
985		In case of problems, replace it with the old file spill.sml
986
987		The current compiler runs into some serious performance problems when
988		constructing a large record.  This can happen when we try to compile a
989		structure with many items.  Even a very simple structure like the following
990		makes the compiler slow down.
991
992		structure Foo = struct
993		val x_1 = 0w1 : Word32.int
994		val x_2 = 0w2 : Word32.int
995		val x_3 = 0w3 : Word32.int
996		...
997		val x_N = 0wN : Word32.int
998		end
999
1000		The following table shows the compile time, from N=1000 to N=4000,
1001		with the old compiler:
1002
1003		N
1004		1000   CPS 100 spill                           0.04u  0.00s  0.00g
1005		MLRISC ra                               0.06u  0.00s  0.05g
1006		(spills = 0 reloads = 0)
1007		TOTAL                                   0.63u  0.07s  0.21g
1008
1009		1100   CPS 100 spill                           8.25u  0.32s  0.64g
1010		MLRISC ra                               5.68u  0.59s  3.93g
1011		(spills = 0 reloads = 0)
1012		TOTAL                                   14.71u  0.99s  4.81g
1013
1014		1500   CPS 100 spill                           58.55u  2.34s  1.74g
1015		MLRISC ra                               5.54u  0.65s  3.91g
1016		(spills = 543 reloads = 1082)
1017		TOTAL                                   65.40u  3.13s  6.00g
1018
1019		2000   CPS 100 spill                           126.69u  4.84s  3.08g
1020		MLRISC ra                               0.80u  0.10s  0.55g
1021		(spills = 42 reloads = 84)
1022		TOTAL                                   129.42u  5.10s  4.13g
1023
1024		3000   CPS 100 spill                           675.59u  19.03s  11.64g
1025		MLRISC ra                               2.69u  0.27s  1.38g
1026		(spills = 62 reloads = 124)
1027		TOTAL                                   682.48u  19.61s  13.99g
1028
1029		4000   CPS 100 spill                           2362.82u  56.28s  43.60g
1030		MLRISC ra                               4.96u  0.27s  2.72g
1031		(spills = 85 reloads = 170)
1032		TOTAL                                   2375.26u  57.21s  48.00g
1033
1034		As you can see the old cps spill module suffers from some serious
1035		performance problem.  But since I cannot decipher the old code fully,
1036		instead of patching the problems up, I'm reimplementing it
1037		with a different algorithm.  The new code is more modular,
1038		smaller when compiled, and substantially faster
1039		(O(n log n) time and O(n) space).  Timing of the new spill module:
1040
1041		4000  CPS 100 spill                           0.02u  0.00s  0.00g
1042		MLRISC ra                               0.25u  0.02s  0.15g
1043		(spills=1 reloads=3)
1044		TOTAL                                   7.74u  0.34s  1.62g
1045
1046		Implementation details:
1047
1048		As far as I can tell, the purpose of the CPS spill module is to make sure the
1049		number of live variables at any program point (the bandwidth)
1050		does not exceed a certain limit, which is determined by the
1051		size of the spill area.
1052
1053		When the bandwidth is too large, we decrease the register pressure by
1054		packing live variables into spill records.  How we achieve this is
1055		completely different than what we did in the old code.
1056
1057		First, there is something about the MLRiscGen code generator
1058		that we should be aware of:
1059
1060		o MLRiscGen performs code motion!
1061
1062		In particular, it will move floating point computations and
1063		address computations involving only the heap pointer to
1064		their use sites (if there is only a single use).
1065		What this means is that if we have a CPS record construction
1066		statement
1067
1068		RECORD(k,vl,w,e)
1069
1070		we should never count the new record address w as live if w
1071		has only one use (which is often the case).
1072
1073		We should do something similar to floating point, but the transformation
1074		there is much more complex, so I won't deal with that.
1075
1076		Secondly, there are now two new cps primops at our disposal:
1077
1078		1. rawrecord of record_kind option
1079		This pure operator allocates some uninitialized storage from the heap.
1080		There are two forms:
1081
1082		rawrecord NONE [INT n]  allocates a tagless record of length n
1083		rawrecord (SOME rk) [INT n] allocates a tagged record of length n
1084		and initializes the tag.
1085
1086		2. rawupdate of cty
1087		rawupdate cty (v,i,x)
1088		Assigns to x to the ith component of record v.
1089		The storelist is not updated.
1090
1091		We use these new primops for both spilling and increment record construction.
1092
1093		1. Spilling.
1094
1095		This is implemented with a linear scan algorithm (but generalized
1096		to trees).  The algorithm will create a single spill record at the
1097		beginning of the cps function and use rawupdate to spill to it,
1098		and SELECT or SELp to reload from it.  So both spills and reloads
1099		are fine-grain operations.  In contrast, in the old algorithm
1100		"spills" have to be bundled together in records.
1101
1102		Ideally, we should sink the spill record construction to where
1103		it is needed.  We can even split the spill record into multiple ones
1104		at the places where they are needed.  But CPS is not a good
1105		representation for global code motion, so I'll keep it simple and
1106		am not attempting this.
1107
1108		2. Incremental record construction (aka record splitting).
1109
1110		Long records with many component values which are simulatenously live
1111		(recall that single use record addresses are not considered to
1112		be live) are constructed with rawrecord and rawupdate.
1113		We allocate space on the heap with rawrecord first, then gradually
1114		fill it in with rawupdate.  This is the technique suggested to me
1115		by Matthias.
1116
1117		Some restrictions on when this is applicable:
1118		1. It is not a VECTOR record.  The code generator currently does not handle
1119		this case. VECTOR record uses double indirection like arrays.
1120		2. All the record component values are defined in the same "basic block"
1121		as the record constructor.  This is to prevent speculative
1122		record construction.
1123
1124		----------------------------------------------------------------------
1125		Name: Allen Leung
1126		Date: 2002/02/22 01:02:00 EST
1127		Tag: leunga-20020222-mlrisc-tools
1128
1129		Minor bug fixes in the parser and rewriter
1130
1131		----------------------------------------------------------------------
1132		Name: Allen Leung
1133		Date: 2002/02/21 20:20:00 EST
1134		Tag: leunga-20020221-peephole
1135
1136		Regenerated the peephole files.  Some contained typos in the specification
1137		and some didn't compile because of pretty printing bugs in the old version
1138		of 'nowhere'.
1139
1140		----------------------------------------------------------------------
1141		Name: Allen Leung
1142		Date: 2002/02/19 20:20:00 EST
1143		Tag: leunga-20020219-mlrisc-tools
1144		Description:
1145
1146		Minor bug fixes to the mlrisc-tools library:
1147
1148		1.  Fixed up parsing colon suffixed keywords
1149		2.  Added the ability to shut the error messages up
1150		3.  Reimplemented the pretty printer and fixed up/improved
1151		the pretty printing of handle and -> types.
1152		4.  Fixed up generation of literal symbols in the nowhere tool.
1153		5.  Added some SML keywords to to sml.sty
1154
1155		----------------------------------------------------------------------
1156		Name: Matthias Blume
1157		Date: 2002/02/19 16:20:00 EST
1158		Tag: blume-20020219-cmffi
1159		Description:
1160
1161		A wild mix of changes, some minor, some major:
1162
1163		* All C FFI-related libraries are now anchored under $c:
1164		$/c.cm      --> $c/c.cm
1165		$/c-int.cm  --> $c/internals/c-int.cm
1166		$/memory.cm --> $c/memory/memory.cm
1167
1168		* "make" tool (in CM) now treats its argument pathname slightly
1169		differently:
1170		1. If the native expansion is an absolute name, then before invoking
1171		the "make" command on it, CM will apply OS.Path.mkRelative
1172		(with relativeTo = OS.FileSys.getDir()) to it.
1173		2. The argument will be passed through to subsequent phases of CM
1174		processing without "going native".  In particular, if the argument
1175		was an anchored path, then "make" will not lose track of that anchor.
1176
1177		* Compiler backends now "know" their respective C calling conventions
1178		instead of having to be told about it by ml-nlffigen.  This relieves
1179		ml-nlffigen from one of its burdens.
1180
1181		* The X86Backend has been split into X86CCallBackend and X86StdCallBackend.
1182
1183		* Export C_DEBUG and C_Debug from $c/c.cm.
1184
1185		* C type encoding in ml-nlffi-lib has been improved to model the conceptual
1186		subtyping relationship between incomplete pointers and their complete
1187		counterparts.  For this, ('t, 'c) ptr has been changed to 'o ptr --
1188		with the convention of instantiating 'o with ('t, 'c) obj whenever
1189		the pointer target type is complete.  In the incomplete case, 'o
1190		will be instantiated with some "'c iobj" -- a type obtained by
1191		using one of the functors PointerToIncompleteType or PointerToCompleteType.
1192
1193		Operations that work on both incomplete and complete pointer types are
1194		typed as taking an 'o ptr while operations that require the target to
1195		be known are typed as taking some ('t, 'c) obj ptr.
1196
1197		voidptr is now a bit "more concrete", namely "type voidptr = void ptr'"
1198		where void is an eqtype without any values.  This makes it possible
1199		to work on voidptr values using functions meant to operate on light
1200		incomplete pointers.
1201
1202		* As a result of the above, signature POINTER_TO_INCOMPLETE_TYPE has
1203		been vastly simplified.
1204
1205		----------------------------------------------------------------------
1206		Name: Matthias Blume
1207		Date: 2002/02/19 10:48:00 EST
1208		Tag: blume-20020219-pqfix
1209		Description:
1210
1211		Applied Chris Okasaki's bug fix for priority queues.
1212
1213		----------------------------------------------------------------------
1214		Name: Matthias Blume
1215		Date: 2002/02/15 17:05:00
1216		Tag: Release_110_39
1217		Description:
1218
1219		Last-minute retagging is becoming a tradition... :-(
1220
1221		This is the working release 110.39.
1222
1223		----------------------------------------------------------------------
1224		Name: Matthias Blume
1225		Date: 2002/02/15 16:00:00 EST
1226		Tag: Release_110_39-orig
1227		Description:
1228
1229		Working release 110.39.  New bootfiles.
1230
1231		(Update: There was a small bug in the installer so it wouldn't work
1232		with all shells.  So I retagged. -Matthias)
1233
1234		----------------------------------------------------------------------
1235		Name: Matthias Blume
1236		Date: 2002/02/15 14:17:00 EST
1237		Tag: blume-20020215-showbindings
1238		Description:
1239
1240		Added EnvRef.listBoundSymbols and CM.State.showBindings.  Especially
1241		the latter can be useful for exploring what bindings are available at
1242		the interactive prompt.  (The first function returns only the list
1243		of symbols that are really bound, the second prints those but also the
1244		ones that CM's autoloading mechanism knows about.)
1245
1246		----------------------------------------------------------------------
1247		Name: Matthias Blume
1248		Date: 2002/02/15 12:08:00 EST
1249		Tag: blume-20020215-iptrs
1250		Description:
1251
1252		Two improvements to ml-nlffigen:
1253
1254		1. Write files only if they do not exist or if their current contents
1255		do not coincide with what's being written.  (That is, avoid messing
1256		with the time stamps unless absolutely necessary.)
1257
1258		2. Implement a "repository" mechanism for generated files related
1259		to "incomplete pointer types".   See the README file for details.
1260
1261		----------------------------------------------------------------------
1262		Name: Matthias Blume
1263		Date: 2002/02/14 11:50:00 EST
1264		Tag: blume-20020214-quote
1265		Description:
1266
1267		Added a type 't t_' to tag.sml (in ml-nlffi-lib.cm).  This is required
1268		because of the new and improved tag generation scheme.  (Thanks to Allen
1269		Leung for pointing it out.)
1270
1271		----------------------------------------------------------------------
1272		Name: Lal George
1273		Date: 2002/02/14 09:55:27 EST 2002
1274		Tag: george-20020214-isabelle-bug
1275		Description:
1276
1277		Fixed the MLRISC bug sent by Markus Wenzel regarding the compilation
1278		of Isabelle on the x86.
1279
1280		From Allen:
1281		-----------
1282		I've found the problem:
1283
1284		in ra-core.sml, I use the counter "blocked" to keep track of the
1285		true number of elements in the freeze queue.  When the counter goes
1286		to zero, I skip examining the queue.  But I've messed up the
1287		bookkeeping in combine():
1288
1289		else ();
1290		case !ucol of
1291		PSEUDO => (if !cntv > 0 then
1292		(if !cntu > 0 then blocked := !blocked - 1 else ();
1293		^^^^^^^^^^^^^^^^^^^^^^^
1294		moveu := mergeMoveList(!movev, !moveu)
1295		)
1296		else ();
1297
1298		combine() is called to coalesce two nodes u and v.
1299		I think I was thinking that if the move counts of u and v are both
1300		greater than zero then after they are coalesced then one node is
1301		removed from the freeze queue.  Apparently I was thinking that
1302		both u and v are of low degree, but that's clearly not necessarily true.
1303
1304
1305		02/12/2002:
1306		Here's the patch.  HOL now compiles.
1307
1308		I don't know how this impact on performance (compile
1309		time or runtime).  This bug caused the RA (especially on the x86)
1310		to go thru the potential spill phase when there are still nodes on the
1311		freeze queue.
1312
1313
1314
1315
1316		----------------------------------------------------------------------
1317		Name: Matthias Blume
1318		Date: 2002/02/13 22:40:00 EST
1319		Tag: blume-20020213-fptr-rtti
1320		Description:
1321
1322		Fixed a bug in ml-nlffigen that was introduced with one of the previous
1323		updates.
1324
1325		----------------------------------------------------------------------
1326		Name: Matthias Blume
1327		Date: 2002/02/13 16:41:00 EST
1328		Tag: blume-20020213-cmlpq
1329		Description:
1330
1331		Added new priority queue export symbols (which have just been added to
1332		smlnj-lib.cm) to CML's version of smlnj-lib.cm.  (Otherwise CML would
1333		not compile and the installer would choke.)
1334
1335		----------------------------------------------------------------------
1336		Name: Matthias Blume
1337		Date: 2002/02/13 16:15:00 EST
1338		Tag: blume-20020213-various
1339		Description:
1340
1341		1. More tweaks to ml-nlffigen:
1342
1343		- better internal datastructures (resulting in slight speedup)
1344		- "-match" option requires exact match
1345		- "localized" gensym counters (untagged structs/unions nested within
1346		other structs/unions or within typedefs get a fresh counter; their
1347		tag will be prefixed by a concatenation of their parents' tags)
1348		- bug fixes (related to calculation of transitive closure of types
1349		to be included in the output)
1350
1351		2. Minor Basis updates:
1352
1353		- added implementations for List.collate and Option.app
1354
1355		----------------------------------------------------------------------
1356		Name: Matthias Blume
1357		Date: 2002/02/11 15:55:00 EST
1358		Tag: blume-20020211-gensym
1359		Description:
1360
1361		Added a "-gensym" option to command line of ml-nlffigen.  This can be
1362		used to specify a "stem" -- a string that is inserted in all "gensym'd"
1363		names (ML structure names that correspond to unnamed C structs, unions,
1364		and enums), so that separate runs of ml-nlffigen do not clash.
1365
1366		----------------------------------------------------------------------
1367		Name: Matthias Blume
1368		Date: 2002/02/11 12:05:00 EST
1369		Tag: blume-20020211-gensml
1370		Description:
1371
1372		A quick fix for a problem with GenSML (in the pgraph-util library):
1373		Make generation of toplevel "local" optional.  (Strictly speaking,
1374		signature definitions within "local" are not legal SML.)
1375
1376		Other than that: updates to INSTALL and cm/TODO.
1377
1378		----------------------------------------------------------------------
1379		Name: Matthias Blume
1380		Date: 2002/02/08 15:00:00 EST
1381		Tag: blume-20020208-uniquepid
1382		Description:
1383
1384		0. Version number has been bumped to 110.38.1.  NEW BOOTFILES!!!
1385
1386		1. The installer (config/install.sh) has gotten smarter:
1387
1388		- Configuration options are a bit easier to specify now
1389		(in config/targets).
1390		- Bug in recognizing .tar.bz2 files fixed.
1391		- Installer automatically resolves dependencies between
1392		configuration options (e.g., if you ask for eXene, you will
1393		also get cml -- regardless whether you asked for it or not).
1394		- Installer can run in "quieter mode" by setting the environment
1395		variable INSTALL_QUIETLY to "true".  "Quieter" does not mean
1396		"completely silent", though.
1397		- Build HashCons library as part of smlnj-lib.
1398
1399		2. A new scheme for assigning persistent identifiers to compilation
1400		units (and, by extension, to types etc.) has been put into place.
1401		This fixes a long-standing bug where types and even dynamic values
1402		can get internally confused, thereby compromising type safety
1403		(abstraction) and dynamic correctness.  See
1404
1405		http://cm.bell-labs.com/cm/cs/who/blume/pid-confusion.tgz
1406
1407		for an example of how things could go wrong until now.
1408
1409		The downside of the new scheme is that pids are not quite as
1410		persistent as they used to be: CM will generate a fresh pid
1411		for every compilation unit that it thinks it sees for the first
1412		time.  That means that if you compile starting from a clean, fresh
1413		source tree at two different times, you end up with different
1414		binaries.
1415
1416		Cutoff recompilation, however, has not been compromised because
1417		CM keeps pid information in special caches between runs.
1418
1419		----------------------------------------------------------------------
1420		Name: Lal George
1421		Date: 2002/02/07 15:34:13 EST 2002
1422		Tag: <none>
1423		Description:
1424
1425		Compilers that generate assembly code may produce  global labels
1426		whose value is resolved at link time. The various peephole optimization
1427		modules did not take this in account.
1428
1429		TODO. The Labels.addrOf function should really return an option
1430		type so that clients are forced to deal with this issue, rather
1431		than an exception being raised.
1432
1433		----------------------------------------------------------------------
1434		Name: Lal George
1435		Date: 2002/02/06 13:55:02 EST
1436		Tag: george-20020206-ra-breakup
1437		Description:
1438
1439		1. A bug fix from Allen.
1440
1441		A typo causes extra fstp %st(0)'s to be generated at compensation
1442		edges, which might cause stack underflow traps at runtime.  This
1443		occurs in fft where there are extraneous fstps right before the 'into'
1444		trap instruction (in this case they are harmless since none of the
1445		integers overflow.)
1446
1447		2. Pulled out various utility modules that were embedded in the modules
1448		of the register allocator. I need these modules for other purposes, but
1449		they are not complete enough to put into a library (just yet).
1450		----------------------------------------------------------------------
1451		Name: Matthias Blume
1452		Date: 2002/01/31 16:05:00 EST
1453		Tag: blume-20020131-sparc-ccalls
1454		Description:
1455
1456		1. C-calls on Sparc needlessly allocated a huge chunk (96 bytes)
1457		of extra stack space by mistake.  Fixed.
1458
1459		2. Bug in logic of handling of command-line options in ml-nlffigen fixed.
1460
1461		----------------------------------------------------------------------
1462		Name: Allen Leung
1463		Date: 2002/01/30
1464		Tag: leunga-20020130-nowhere-bug-fix
1465		Description:
1466
1467		MLRISC bug fixes:
1468		1. Fixed a bindings computation bug in the 'nowhere' program generator tool.
1469		2. MachineInt.fromString was negating its value.
1470
1471		----------------------------------------------------------------------
1472		Name: Matthias Blume
1473		Date: 2002/01/29
1474		Tag: blume-20020129-INSTALL
1475		Description:
1476
1477		- Added somewhat detailed installation instructions (file INSTALL).
1478		- Fixed curl-detection bug in config/install.sh.
1479		- It is now possible to select the URL getter using the URLGETTER
1480		environment variable:
1481
1482		not set / "unknown"      --> automatic detection (script tries wget,
1483		curl, and lynx)
1484		"wget" / "curl" / "lynx" --> use the specified program (script "knows"
1485		how to properly invoke them)
1486		other                    --> use $URLGETTER directly, it must take
1487		precisely two command-line arguments
1488		(source URL and destination file name)
1489
1490		----------------------------------------------------------------------
1491		Name: Matthias Blume
1492		Date: 2002/01/28
1493		Tag: blume-20020128-sparc-ccalls
1494		Description:
1495
1496		- Fixed problem with calculation of "used" registers in sparc-c-calls.
1497		- Make use of the allocParam argument in sparc-c-calls.
1498
1499		----------------------------------------------------------------------
1500		Name: Matthias Blume
1501		Date: 2002/01/28
1502		Tag: blume-20020128-allocParam
1503		Description:
1504
1505		John Reppy:  Changes c-calls API to accept client-callback for
1506		allocating extra stack space.
1507		me: Corresponding changes to mlriscGen (using a dummy argument that
1508		does not change the current behavior).
1509
1510		----------------------------------------------------------------------
1511		Name: Matthias Blume
1512		Date: 2002/01/28 12:00:00
1513		Tag: Release_110_38
1514		Description:
1515
1516		This time for real!!!
1517
1518		----------------------------------------------------------------------
1519		Name: Matthias Blume
1520		Date: 2002/01/28 10:56:00 EST
1521		Tag: blume-20020128-retraction
1522		Description:
1523
1524		0. Retracted earlier 110.38.  (The Release_110_38 tag has been replaced
1525		with blume-Release_110_38-retracted.)
1526
1527		1. Fixed a problem with incorrect rounding modes in real64.sml.
1528		(Thanks to Andrew Mccreight <andrew.mccreight@yale.edu>.)
1529
1530		2. A bug in ml-nlffigen related to the handling of unnamed structs, unions,
1531		and enums fixed.  The naming of corresponding ML identifiers should
1532		now be consistent again.
1533
1534		----------------------------------------------------------------------
1535		Name: Allen Leung
1536		Date: 2002/01/27
1537		Tag: leunga-20020127-nowhere
1538		Description:
1539
1540		Added a target called nowhere in the configuration scripts.
1541		Enabling this will build the MLRISC 'nowhere' tool (for translating
1542		programs with where-clauses into legal SML code) during installation.
1543
1544		----------------------------------------------------------------------
1545		Name: Matthias Blume
1546		Date: 2002/01/25 21:27:00 EST
1547		Tag: blume-Release_110_38-retracted
1548		Description:
1549
1550		Call it a (working) release!  Version is 110.38. Bootfiles are ready.
1551
1552		README will be added later.
1553
1554		!!! NOTE:  Re-tagged as blume-Release_110_38-retracted. Original tag
1555		(Release_110_38) removed.  Reason: Last-minute bug fixes.
1556
1557		----------------------------------------------------------------------
1558		Name: Matthias Blume
1559		Date: 2002/01/25
1560		Tag: blume-20020125-ffi
1561		Description:
1562
1563		A large number of tweaks and improvements to ml-nlffi-lib and
1564		ml-nlffigen:
1565
1566		- ML represenation types have been streamlined
1567		- getter and setter functions work with concrete values, not abstract
1568		ones where possible
1569		- ml-nlffigen command line more flexible (see README file there)
1570		- some bugs have been fixed (hopefully)
1571
1572		----------------------------------------------------------------------
1573		Name: Lal George
1574		Date: 2002/01/24
1575		Tag: george-20020124-risc-ra-interface
1576		Description:
1577
1578		There is a dramatic simplification in the interface to the
1579		register allocator for RISC architectures as a result of making
1580		parallel copy instructions explicit.
1581
1582		----------------------------------------------------------------------
1583		Name: Matthias Blume
1584		Date: 2002/01/22
1585		Tag: blume-20020122-x86-ccalls
1586		Description:
1587
1588		Bug fix for c-calls on x86 (having to do with how char- and
1589		short-arguments are being handled).
1590
1591		----------------------------------------------------------------------
1592		Name: Matthias Blume
1593		Date: 2002/01/21
1594		Tag: blume-20020121-ff
1595		Description:
1596
1597		Another day of fiddling with the FFI...
1598
1599		1. Bug fix/workaround:  CKIT does not complain about negative array
1600		dimensions, so ml-nlffigen has to guard itself against this possibility.
1601		(Otherwise a negative dimension would send it into an infinite loop.)
1602
1603		2. Some of the abstract types (light objects, light pointers, most "base"
1604		types) in structure C are now eqtypes.
1605
1606		3. Added constructors and test functions for NULL function pointers.
1607
1608		----------------------------------------------------------------------
1609		Name: Matthias Blume
1610		Date: 2002/01/18
1611		Tag: blume-20020118-ready-for-new-release
1612		Description:
1613
1614		Made config/srcarchiveurl point to a new place.  (Will provide boot
1615		files shortly.)
1616
1617		Maybe we christen this to be 110.38?
1618
1619		----------------------------------------------------------------------
1620		Name: Matthias Blume
1621		Date: 2002/01/18
1622		Tag: blume-20020118-more-ffifiddle
1623		Description:
1624
1625		Today's FFI fiddling:
1626
1627		- Provided a structure CGetSet with "convenient" versions of C.Get.* and
1628		C.Set.* that use concrete (MLRep.*) arguments and results instead
1629		of abstract ones.
1630
1631		- Provided word-style bit operations etc. for "int" representation
1632		types in MLRep.S<Foo>Bitops where <Foo> ranges over Char, Int, Short,
1633		and Long.
1634
1635		----------------------------------------------------------------------
1636		Name: Matthias Blume
1637		Date: 2002/01/18
1638		Tag: blume-20020118-use-x86-fp
1639		Description:
1640
1641		Now that x86-fast-fp seems to be working, I turned it back on again
1642		by default.  (Seems to work fine now, even with the FFI.)
1643
1644		Other than that, I added some documentation about the FFI to
1645		src/ml-nlffigen/README and updated the FFI test examples in
1646		src/ml-nlffi-lib/Tests/*.
1647
1648		----------------------------------------------------------------------
1649		Name: Allen Leung
1650		Date: 2002/01/17
1651		Tag: leunga-20020117-x86-fast-fp-call
1652		Description:
1653
1654		1. Fixed a problem with handling return fp values when x86's fast fp
1655		mode is turned on.
1656
1657		2. Minor pretty printing fix for cellset.  Print %st(0) as %st(0) instead
1658		of %f32.
1659
1660		3. Added a constructor INT32lit to the ast of MLRISC tools.
1661
1662		----------------------------------------------------------------------
1663		Name: Matthias Blume
1664		Date: 2002/01/16
1665		Tag: blume-20020116-ffifiddle
1666		Description:
1667
1668		More fiddling with the FFI interface:
1669
1670		- Make constness 'c instead of rw wherever possible.  This eliminates
1671		the need for certain explicit coercions.  (However, due to ML's
1672		value polymorphism, there will still be many cases where explicit
1673		coercions are necessary.  Phantom types are not the whole answer
1674		to modeling a subtyping relationship in ML.)
1675
1676		- ro/rw coersions for pointers added.  (Avoids the detour through */&.)
1677
1678		- "printf" test example added to src/ml-nlffi-lib/Tests.  (Demonstrates
1679		clumsy workaround for varargs problem.)
1680
1681		----------------------------------------------------------------------
1682		Name: Lal George
1683		Date: 2002/01/15
1684		Tag: <none>
1685		Description:
1686
1687		1. Since COPY instructions are no longer native to the architecture,
1688		a generic functor can be used to implement the expandCopies function.
1689
1690		2. Allowed EXPORT and IMPORT pseudo-op declarations to appear inside a
1691		TEXT segment.
1692
1693		----------------------------------------------------------------------
1694		Name: Matthias Blume
1695		Date: 2002/01/15
1696		Tag: blume-20020115-ffiupdates
1697		Description:
1698
1699		1. Fix for bug resulting in single-precision float values being returned
1700		incorrectly from FFI calls.
1701
1702		2. Small modifications to C FFI API:
1703
1704		- memory-allocation routines return straight objects (no options)
1705		and raise an exception in out-of-memory situations
1706		- unsafe extensions to cast between function pointers and pointers
1707		from/to ints
1708		- added structure C_Debug as an alternative to structure C where
1709		pointer-dereferencing (|*| and |*!) always check for null-pointers
1710		- added open_lib' to DynLinkage;  open_lib' works like open_lib
1711		but also takes a (possibly empty) list of existing library handles
1712		that the current library depends on
1713
1714		----------------------------------------------------------------------
1715		Name: Matthias Blume
1716		Date: 2002/01/10
1717		Tag: blume-20020110-newffigen
1718		Description:
1719
1720		1. Updates to portable graph code.
1721
1722		2. Major update to ml-nlffigen and ml-nlffi-lib.  Things are much
1723		more scalable now so that even huge interfaces such as the one
1724		for GTK compile in finite time and space. :-)
1725		See src/ml-nlffigen/README for details on what's new.
1726
1727		----------------------------------------------------------------------
1728		Name: Lal George
1729		Date: 2001/01/09 14:31:35 EST 2002
1730		Tag: george-20011206-rm-native-copy
1731		Description:
1732
1733		Removed the native COPY and FCOPY instructions
1734		from all the architectures and replaced it with the
1735		explicit COPY instruction from the previous commit.
1736
1737		It is now possible to simplify many of the optimizations
1738		modules that manipulate copies. This has not been
1739		done in this change.
1740
1741		----------------------------------------------------------------------
1742		Name: Lal George
1743		Date: 2001/12/06 16:50:13 EST 2001
1744		Tag: george-20011206-mlrisc-instruction
1745		Description:
1746
1747		Changed the representation of instructions from being fully abstract
1748		to being partially concrete. That is to say:
1749
1750		from
1751		type instruction
1752
1753		to
1754		type instr                              (* machine instruction *)
1755
1756		datatype instruction =
1757		LIVE of {regs: C.cellset, spilled: C.cellset}
1758		| KILL of {regs: C.cellset, spilled: C.cellset}
1759		| COPYXXX of {k: CB.cellkind, dst: CB.cell list, src: CB.cell list}
1760		| ANNOTATION of {i: instruction, a: Annotations.annotation}
1761		| INSTR of instr
1762
1763		This makes the handling of certain special instructions that appear on
1764		all architectures easier and uniform.
1765
1766		LIVE and KILL say that a list of registers are live or killed at the
1767		program point where they appear. No spill code is generated when an
1768		element of the 'regs' field is spilled, but the register is moved to
1769		the 'spilled' (which is present, more for debugging than anything else).
1770
1771		LIVE replaces the (now deprecated) DEFFREG instruction on the alpha.
1772		We used to generate:
1773
1774		DEFFREG f1
1775		f1 := f2 + f3
1776		trapb
1777
1778		but now generate:
1779
1780		f1 := f2 + f3
1781		trapb
1782		LIVE {regs=[f1,f2,f3], spilled=[]}
1783
1784		Furthermore, the DEFFREG (hack) required that all floating point instruction
1785		use all registers mentioned in the instruction. Therefore f1 := f2 + f3,
1786		defines f1 and uses [f1,f2,f3]! This hack is no longer required resulting
1787		in a cleaner alpha implementation. (Hopefully, intel will not get rid of
1788		this architecture).
1789
1790		COPYXXX is intended to replace the parallel COPY and FCOPY  available on
1791		all the architectures. This will result in further simplification of the
1792		register allocator that must be aware of them for coalescing purposes, and
1793		will also simplify certain aspects of the machine description that provides
1794		callbacks related to parallel copies.
1795
1796		ANNOTATION should be obvious, and now INSTR represents the honest to God
1797		machine instruction set!
1798
1799		The <arch>/instructions/<arch>Instr.sml files define certain utility
1800		functions for making porting easier -- essentially converting upper case
1801		to lower case. All machine instructions (of type instr) are in upper case,
1802		and the lower case form generates an MLRISC instruction. For example on
1803		the alpha we have:
1804
1805		datatype instr =
1806		LDA of {r:cell, b:cell, d:operand}
1807		| ...
1808
1809		val lda : {r:cell, b:cell, d:operand} -> instruction
1810		...
1811
1812		where lda is just (INSTR o LDA), etc.
1813
1814		----------------------------------------------------------------------
1815		Name: Matthias Blume
1816		Date: 2001/11/22 21:40:00 EST
1817		Tag: Release_110_37
1818		Description:
1819
1820		Release 110.37.  This time for real.
1821
1822		----------------------------------------------------------------------
1823		Name: Matthias Blume
1824		Date: 2001/11/21 16:35:00 EST
1825		Tag: blume-20011121-foot-in-mouth
1826		Description:
1827
1828		Removed the "Release_110_37" tag because of a serious bug.
1829		This will be re-tagged once the bug is fixed.
1830
1831		----------------------------------------------------------------------
1832		Name: Matthias Blume
1833		Date: 2001/11/21 16:14:00 EST
1834		Tag: blume-20011121-forgottenfile
1835		Description:
1836
1837		Forgot to add a file.  (Just a .tex-file -- part of
1838		the CM manual source.)
1839
1840		----------------------------------------------------------------------
1841		Name: Matthias Blume
1842		Date: 2001/11/21 16:10:00 EST
1843		Tag: blume-20011121-invalid_110_37
1844		Description:
1845
1846		Note: I removed the original tag "Release_110_37" from this commit
1847		because we found a serious bug in all non-x86 backends.
1848		- Matthias
1849
1850		1. Modifications to the SML/NJ code generator and to the runtime system
1851		so that code object name strings are directly inserted into code
1852		objects at code generation time.  The only business the runtime system
1853		has with this is now to read the name strings on occasions.
1854		(The encoding of the name string has also changed somewhat.)
1855
1856		2. CM now implements a simple "set calculus" for specifying export lists.
1857		In particular, it is now possible to refer to the export lists of
1858		other libraries/groups/sources and form unions as well as differences.
1859		See the latest CM manual for details.
1860
1861		3. An separate notion of "proxy" libraries has again be eliminated from
1862		CM's model.  (Proxy libraries are now simply a special case of using
1863		the export list calculus.)
1864
1865		4. Some of the existing libraries now take advantage of the new set
1866		calculus.
1867		(Notice that not all libraries have been converted because some
1868		of the existing .cm-files are supposed to be backward compatible
1869		with 110.0.x.)
1870
1871		5. Some cleanup in stand-alone programs.  (Don't use "exnMessage" -- use
1872		"General.exnMessage"!  The former relies on a certain hook to be
1873		initialized, and that often does not happen in the stand-alone case.)
1874
1875		----------------------------------------------------------------------
1876		Name: Lal George
1877		Date: 2001/11/21  13:56:18 EST
1878		Tag: george-2001121-pseudo-ops
1879		Description:
1880
1881		Implemented a complete redesign of MLRISC pseudo-ops. Now there
1882		ought to never be any question of incompatabilities with
1883		pseudo-op syntax expected by host assemblers.
1884
1885		For now, only modules supporting GAS syntax are implemented
1886		but more should follow, such as MASM, and vendor assembler
1887		syntax, e.g. IBM as, Sun as, etc.
1888
1889		----------------------------------------------------------------------
1890		Name: Matthias Blume
1891		Date: 2001/11/14 11:52:00 EST
1892		Tag: blume-20011114-srcname
1893		Description:
1894
1895		1. Routed the name of the current source file to mlriscgen where it
1896		should be directly emitted into the code object.  (This last part
1897		is yet to be done.)
1898
1899		2. Some cleanup of the pgraph code to make it match the proposal that
1900		I put out the other day.  (The proposal notwithstanding, things are
1901		still in flux here.)
1902
1903		----------------------------------------------------------------------
1904		Name: Lal George
1905		Date: 2001/11/14 09:44:04 EST
1906		Tag:
1907		Description:
1908
1909		Fix for a backpatching bug reported by Allen.
1910
1911		Because the boundary between short and long span-dependent
1912		instructions is +/- 128, there are an astounding number of
1913		span-dependent instructions whose size is over estimated.
1914
1915		Allen came up with the idea of letting the size of span
1916		dependent instructions be non-monotonic, for a maxIter
1917		number of times, after which the size must be monotonically
1918		increasing.
1919
1920		This table shows the number of span-dependent instructions
1921		whose size was over-estimated as a function of maxIter, for the
1922		file Parse/parse/ml.grm.sml:
1923
1924		maxIter            # of instructions:
1925		10                      687
1926		20                      438
1927		30                      198
1928		40                        0
1929
1930		In compiling the compiler, there is no significant difference in
1931		compilation speed between maxIter=10 and maxIter=40. Actually,
1932		my measurements showed that maxIter=40 was a tad faster than
1933		maxIter=10! Also 96% of the  files in the compiler reach a fix
1934		point within 13 iterations, so fixing maxIter at 40, while high,
1935		is okay.
1936
1937		----------------------------------------------------------------------
1938		Name: Matthias Blume
1939		Date: 2001/10/31 15:25:00 EST
1940		Tag: blume-20011031-pgraph
1941		Description:
1942
1943		CKIT:
1944		* Changed the "Function" constructor of type Ast.ctype to carry optional
1945		argument identifiers.
1946		* Changed the return type of TypeUtil.getFunction accordingly.
1947		* Type equality ignores the argument names.
1948		* TypeUtil.composite tries to preserve argument names but gives up quickly
1949		if there is a mismatch.
1950
1951		installation script:
1952		* attempts to use "curl" if available (unless "wget" is available as well)
1953
1954		CM:
1955		* has an experimental implementation of "portable graphs" which I will
1956		soon propose as an implementation-independent library format
1957		* there are also new libraries $/pgraph.cm and $/pgraph-util.cm
1958
1959		NLFFI-LIB:
1960		* some cleanup (all cosmetic)
1961
1962		NLFFIGEN:
1963		* temporarily disabled the mechanism that suppresses ML output for
1964		C definitions whose identifiers start with an underscore character
1965		* generate val bindings for enum constants
1966		* user can request that only one style (light or heavy) is being used;
1967		default is to use both (command-line arguments: -heavy and -light)
1968		* fixed bug in handling of function types involving incomplete pointers
1969		* generate ML entry points that take record arguments (i.e., using
1970		named arguments) for C functions that have a prototype with named
1971		arguments
1972		(see changes to CKIT)
1973
1974		----------------------------------------------------------------------
1975		Name: Allen Leung
1976		Date: 2001/10/27 20:34:00 EDT
1977		Tag: leunga-20011027-x86-fast-fp-call
1978		Description:
1979
1980		Fixed the bug described in blume-20010920-slowfp.
1981
1982		The fix involves
1983		1. generating FCOPYs in FSTP in ia32-svid
1984		2. marking a CALL with the appropriate annotation
1985
1986		----------------------------------------------------------------------
1987		Name: Matthias Blume
1988		Date: 2001/10/16 11:32:00 EDT
1989		Tag: blume-20011016-netbsd
1990		Description:
1991
1992		Underscore patch from Chris Richards (fixing problem with compiling
1993		runtime system under recent NetBSD).
1994
1995		----------------------------------------------------------------------
1996		Name: Allen Leung
1997		Date: 2001/10/12 17:18:32 EDT 2001
1998		Tag: leung-20011012-x86-printflowgraph
1999		Description:
2000
2001		X86RA now uses a valid (instead of dummy) PrintFlowgraph module.
2002
2003		----------------------------------------------------------------------
2004		Name: Lal George
2005		Date: 2001/10/11 23:51:34 EDT
2006		Tag: george-20011011-too-many-instrs
2007		Description:
2008
2009		The representation of a program point never expected to see more
2010		than 65536 instructions in a basic block!
2011
2012		----------------------------------------------------------------------
2013		Name: Lal George
2014		Date: 2001/10/09 09:41:37 EDT
2015		Tag: george-20011008-mlrisc-labels
2016		Description:
2017
2018		Changed the machine description files to support printing of
2019		local and global labels in assembly code, based on host assembler
2020		conventions.
2021
2022		----------------------------------------------------------------------
2023		Name: Matthias Blume
2024		Date: 2001/09/25 15:25:00 EDT
2025		Tag: blume-20010925-exninfo
2026		Description:
2027
2028		I provided a non-hook implementation of exnName (at the toplevel) and
2029		made the "dummy" implementation of exnMessage (at the toplevel) more
2030		useful: if nothing gets "hooked in", then at least you are going to
2031		see the exception name and a message indicating why you don't see more.
2032
2033		[For the time being, programs that need exnMessage and want to use
2034		ml-build should either use General.exnMessage (strongly recommended) or
2035		refer to structure General at some other point so that CM sees a
2036		static dependency.]
2037
2038		[Similar remarks go for "print" and "use":  If you want to use their
2039		functionality in stand-alone programs generated by ml-build, then use
2040		TextIO.output and Backend.Interact.useFile (from $smlnj/compiler.cm).]
2041
2042		----------------------------------------------------------------------
2043		Name: Matthias Blume
2044		Date: 2001/09/20 17:28:00 EDT
2045		Tag: blume-20010920-slowfp
2046		Description:
2047
2048		Allen says that x86-fast-fp is not safe yet, so I turned it off again...
2049
2050		----------------------------------------------------------------------
2051		Name: Matthias Blume
2052		Date: 2001/09/20 17:20:00 EDT
2053		Tag: blume-20010920-canonicalpaths
2054		Description:
2055
2056		0. Updated the BOOT file (something that I forgot to do earlier).
2057
2058		1. Small internal change to CM so that it avoids "/../" in filenames
2059		as much as possible (but only where it is safe).
2060
2061		2. Changed config/_run-sml (resulting in a changed bin/.run-sml) so
2062		that arguments that contain delimiters are passed through correctly.
2063		This change also means that all "special" arguments of the form
2064		@SMLxxx... must come first.
2065
2066		3. Changed install script to put relative anchor names for tool commands
2067		into pathconfig.
2068
2069		----------------------------------------------------------------------
2070		Name: Matthias Blume
2071		Date: 2001/09/18 15:35:00 EDT
2072		Tag: blume-20010918-readme11036
2073		Description:
2074
2075		Added README files.
2076
2077		----------------------------------------------------------------------
2078		Name: Matthias Blume
2079		Date: 2001/09/18 11:45:00 EDT
2080		Tag: Release_110_36 (retag)
2081		Description:
2082
2083		Fixed mistake in config/preloads. Retagged as 110.36.
2084
2085		----------------------------------------------------------------------
2086		Name: Matthias Blume
2087		Date: 2001/09/18 09:40:00 EDT
2088		Tag: Release_110_36_orig (tag changed)
2089		Description:
2090
2091		New version (110.36).  New bootfiles.
2092
2093		----------------------------------------------------------------------
2094		Name: Matthias Blume
2095		Date: 2001/09/14 16:15:00 EDT
2096		Tag: blume-20010914-x86fastfp
2097		Description:
2098
2099		John committed some changes that Allen made, in particular a (hopefully)
2100		correctly working version of the x86-fp module.
2101
2102		I changed the default setting of the Control.MLRISC.getFlag "x86-fast-fp"
2103		flag to "true".  Everything seems to compile to a fixpoint ok, and
2104		"mandelbrot" speeds up by about 15%.
2105
2106		----------------------------------------------------------------------
2107		Name: Matthias Blume
2108		Date: 2001/09/13 11:20:00 EDT
2109		Tag: blume-20010913-minimal
2110		Description:
2111
2112		1. Stefan Monnier's patch to fix a miscompilation problem that
2113		was brought to light by John Reppy's work on Moby.
2114
2115		2. Implemented a minimal "structure Compiler" that contains just
2116		"version" and "architecture".  The minimal version will be
2117		available when the full version is not.  This is for backward-
2118		compatibility with code that wants to test Compiler.version.
2119
2120		----------------------------------------------------------------------
2121		Name: Matthias Blume
2122		Date: 2001/08/28 14:03:00 EDT
2123		Tag: blume-20010828-ml-lex
2124		Description:
2125
2126		Fix for bug 1581, received from Neophytos Michael.
2127
2128		----------------------------------------------------------------------
2129		Name: Matthias Blume
2130		Date: 2001/08/27 11:20:00 EDT
2131		Tag: blume-20010827-readme11035
2132		Description:
2133
2134		Fleshed out the README file for 110.35.
2135
2136		----------------------------------------------------------------------
2137		Name: Matthias Blume
2138		Date: 2001/08/24 17:10:00 EDT
2139		Tag: Release_110_35
2140		Description:
2141
2142		New version number (110.35).  New bootfiles.
2143
2144		----------------------------------------------------------------------
2145		Name: Lal George
2146		Date: 2001/08/24 13:47:18 EDT 2001
2147		Tag: george-20010824-MLRISC-graphs
2148		Description:
2149
2150		removed clusters from MLRISC completely and replaced with graphs.
2151
2152		----------------------------------------------------------------------
2153		Name: Matthias Blume
2154		Date: 2001/08/23 17:50:00 EDT
2155		Tag: blume-20010823-toplevel
2156		Description:
2157
2158		- some reorganization of the code that implements various kinds of
2159		environments in the compiler (static, dynamic, symbolic, combined)
2160		- re-implemented the EnvRef module so that evalStream works properly
2161		(if the stream contains references to "use", "CM.make", etc.)
2162		- cleaned up evalloop.sml and interact.sml (but they need more cleaning)
2163
2164		----------------------------------------------------------------------
2165		Name: Matthias Blume
2166		Date: 2001/08/20 15:50 EDT
2167		Tag: blume20010820-slipup
2168		Description:
2169
2170		I forgot to commit a few files.  Here they are...
2171
2172		----------------------------------------------------------------------
2173		Name: Matthias Blume
2174		Date: 2001/08/20 15:35:00 EDT
2175		Tag: blume-20010820-debugprof
2176		Description:
2177
2178		!!!! NEW BOOTFILES !!!!
2179
2180		This is another round of reorganizing the compiler sources.  This
2181		time the main goal was to factor out all the "instrumentation"
2182		passes (for profiling and backtracing) into their own library.
2183		The difficulty was to do it in such a way that it does not depend
2184		on elaborate.cm but only on elabdata.cm.
2185
2186		Therefore there have been further changes to both elaborate.cm and
2187		elabdata.cm -- more "generic" things have been moved from the former
2188		to the latter.  As a result, I was forced to split the assignment
2189		of numbers indicating "primtyc"s into two portions: SML-generic and
2190		SML/NJ-specific.  Since it would have been awkward to maintain,
2191		I bit the bullet and actually _changed_ the mapping between these
2192		numbers and primtycs.  The bottom line of this is that you need
2193		a new set of bin- and bootfiles.
2194
2195		I have built new bootfiles for all architectures, so doing a fresh
2196		checkout and config/install.sh should be all you need.
2197
2198		The newly created library's name is
2199
2200		$smlnj/viscomp/debugprof.cm
2201
2202		and its sources live under
2203
2204		src/compiler/DebugProf
2205
2206		----------------------------------------------------------------------
2207		Name: Matthias Blume
2208		Date: 2001/08/15 17:15:00 EDT
2209		Tag: blume-20010815-compreorg
2210		Description:
2211
2212		This is a first cut at reorganizing the CM libraries that make up the
2213		core of the compiler.  The idea is to separate out pieces that could
2214		be used independently by tools, e.g., the parser, the typechecker, etc.
2215
2216		The current status is a step in this direction, but it is not quite
2217		satisfactory yet.  Expect more changes in the future.
2218
2219		Here is the current (new) organization...
2220
2221		What used to be $smlnj/viscomp/core.cm is now divided into
2222		six CM libraries:
2223
2224		$smlnj/viscomp/basics.cm
2225		/parser.cm
2226		/elabdata.cm
2227		/elaborate.cm
2228		/execute.cm
2229		/core.cm
2230
2231		The CM files for these libraries live under src/system/smlnj/viscomp.
2232		All these libraries are proxy libraries that contain precisely
2233		one CM library component.  Here are the locations of the components
2234		(all within the src/compiler tree):
2235
2236		Basics/basics.cm
2237		Parse/parser.cm
2238		ElabData/elabdata.cm
2239		Elaborator/elaborate.cm
2240		Execution/execute.cm
2241		core.cm
2242
2243		[This organization is the same that has been used already
2244		for a while for the architecture-specific parts of the visible
2245		compiler and for the old version of core.cm.]
2246
2247		As you will notice, many source files have been moved from their
2248		respective original locations to a new home in one of the above
2249		subtrees.
2250
2251		The division of labor between the new libraries is the following:
2252
2253		basics.cm:
2254		- Simple, basic definitions that pertain to many (or all) of
2255		the other libraries.
2256		parser.cm:
2257		- The SML parser, producing output of type Ast.dec.
2258		- The type family for Ast is also defined and exported here.
2259		elabdata.cm:
2260		- The datatypes that describe input and output of the elaborator.
2261		This includes types, absyn, and static environments.
2262		elaborator.cm:
2263		- The SML/NJ type checker and elaborator.
2264		This maps an Ast.dec (with a given static environment) to
2265		an Absyn.dec (with a new static environment).
2266		- This libraries implements certain modules that used to be
2267		structures as functors (to remove dependencies on FLINT).
2268		execute.cm:
2269		- Everything having to do with executing binary code objects.
2270		- Dynamic environments.
2271		core.cm:
2272		- SML/NJ-specific instantiations of the elaborator and MLRISC.
2273		- Top-level modules.
2274		- FLINT (this should eventually become its own library)
2275
2276		Notes:
2277
2278		I am not 100% happy with the way I separated the elaborator (and its
2279		data structures) from FLINT.  Two instances of the same problem:
2280
2281		1. Data structures contain certain fields that carry FLINT-specific
2282		information.  I hacked around this using exn and the property list
2283		module from smlnj-lib.  But the fact that there are middle-end
2284		specific fields around at all is a bit annoying.
2285
2286		2. The elaborator calculates certain FLINT-related information.  I tried
2287		to make this as abstract as I could using functorization, but, again,
2288		the fact that the elaborator has to perform calculations on behalf
2289		of the middle-end at all is not nice.
2290
2291		3. Having to used exn and property lists is unfortunate because it
2292		weakens type checking.  The other alternative (parameterizing
2293		nearly *everything*) is not appealing, though.
2294
2295		I removed the "rebinding =" warning hack because due to the new organization
2296		it was awkward to maintain it.  As a result, the compiler now issues some of
2297		these warnings when compiling init.cmi during bootstrap compilation. On
2298		the plus side, you also get a warning when you do, for example:
2299		val op = = Int32.+
2300		which was not the case up to now.
2301
2302		I placed "assign" and "deref" into the _Core structure so that the
2303		code that deals with the "lazy" keyword can find them there.  This
2304		removes the need for having access to the primitive environment
2305		during elaboration.
2306
2307		----------------------------------------------------------------------
2308		Name: Matthias Blume
2309		Date: 2001/08/13
2310		Tag: blume-20010813-closures
2311		Description:
2312
2313		This fix was sent to us by Zhong Shao.  It is supposed to improve the
2314		performance of certain loops by avoiding needless closure allocation.
2315
2316		----------------------------------------------------------------------
2317		Name: Lal George
2318		Date: 2001/07/31 10:03:23 EDT 2001
2319		Tag: george-20010731-x86-fmalloc
2320		Description: Fixed bug in x86 calls
2321
2322		There was a bug where call instructions would mysteriously
2323		vanish. The call instruction had to be one that returned
2324		a floating point value.
2325
2326		----------------------------------------------------------------------
2327		Name: Lal George
2328		Date: 2001/07/19 16:36:29 EDT 2001
2329		Tag: george-20010719-simple-cells
2330		Description:
2331
2332		I have dramatically simplified the interface for CELLS in MLRISC.
2333
2334		In summary, the cells interface is broken up into three parts:
2335
2336		1. CellsBasis : CELLS_BASIS
2337
2338		CellsBasis is a top level structure and common for all
2339		architectures.  it contains the definitions of basic datatypes
2340		and utility  functions over these types.
2341
2342		2. functor Cells() : CELLS
2343
2344		Cells generates an interface for CELLS that incorporates the
2345		specific resources on the target architecture, such as the
2346		presence of special register classes, their number and size,
2347		and various useful substructures.
2348
2349		3. <ARCH>CELLS
2350
2351		e.g. SparcCells: SPARCCELLS
2352
2353		<ARCH>CELLS usually contains additional bindings for special
2354		registers  on the architecture, such as:
2355
2356		val r0 : cell           (* register zero *)
2357		val y : cell            (* Y register *)
2358		val psr : cell          (* processor status register *)
2359		...
2360
2361		The structure returned by applying the Cells functor is opened
2362		in this interface.
2363
2364		The main implication of all this is that the datatypes for cells is
2365		split between CellsBasis and CELLS -- a fairly simple change for user
2366		code.
2367
2368		In the old scheme the CELLS interface had a definitional binding of
2369		the form:
2370
2371		signature CELLS = sig
2372
2373		structure CellsBasis = CellsBasis
2374
2375		...
2376
2377		end
2378
2379		With all the sharing constraints that goes on in MLRISC, this old
2380		design  quickly leads to errors such as:
2381
2382		"structure definition spec inside of sharing ... "
2383
2384
2385		and appears to require an unacceptable amount of sharing and where
2386		constraint hackery.
2387
2388		I think this error message (the interaction of definitional specs and
2389		sharing) requires more explanation on our web page.
2390
2391		----------------------------------------------------------------------
2392		Name: Matthias Blume
2393		Date: 2001/07/19 15:00:00 EDT
2394		Tag: blume-20010719-libreorg
2395		Description:
2396
2397		This update puts together a fairly extensive but straightforward change
2398		to the way the libraries that implement the interactive system are
2399		organized:
2400
2401		The biggest change is the elimination of structure Compiler.  As a
2402		replacement for this structure, there is now a CM library
2403		(known as $smlnj/compiler.cm or $smlnj/compiler/current.cm)
2404		that exports all the substructures of the original structure Compiler
2405		directly.  So instead of saying Compiler.Foo.bar one now simply
2406		says Foo.bar.  (The CM libraries actually export a collection of
2407		structures that is richer than the collection of substructures of
2408		structure Compiler.)
2409
2410		To make the transition smooth, there is a separate library called
2411		$smlnj/compiler/compiler.cm which puts together and exports the
2412		original structure Compiler (or at least something very close to it).
2413
2414		There are five members of the original structure Compiler
2415		that are not exported directly but which instead became members
2416		of a new structure Backend (described by signature BACKEND).  These are:
2417		structure Profile (: PROFILE), structure Compile (: COMPILE), structure
2418		Interact (: INTERACT), structure Machine (: MACHINE), and val
2419		architecture (: string).
2420
2421		Structure Compiler.Version has become structure CompilerVersion.
2422
2423		Cross-compilers for alpha32, hppa, ppc, sparc, and x86 are provided
2424		by $smlnj/compiler/<arch>.cm where <arch> is alpha32, hppa, ppc, sparc,
2425		or x86, respectively.
2426		Each of these exports the same frontend structures that
2427		$smlnj/compiler.cm exports.  But they do not have a structure Backend
2428		and instead export some structure <Arch>Backend where <Arch> is Alpha32,
2429		Hppa, PPC, Sparc, or X86, respectively.
2430
2431		Library $smlnj/compiler/all.cm exports the union of the exports of
2432		$smlnj/compiler/<arch>.cm
2433
2434		There are no structures <Arch>Compiler anymore, use
2435		$smlnj/compiler/<arch>.cm instead.
2436
2437		Library host-compiler-0.cm is gone.  Instead, the internal library
2438		that instantiates CM is now called cm0.cm.  Selection of the host
2439		compiler (backend) is no longer done here but. (Responsibility for it
2440		now lies with $smlnj/compiler/current.cm.  This seems to be more
2441		logical.)
2442
2443		Many individual files have been moved or renamed.  Some files have
2444		been split into multiple files, and some "dead" files have been deleted.
2445
2446		Aside from these changes to library organization, there are also changes
2447		to the way the code itself is organized:
2448
2449		Structure Binfile has been re-implemented in such a way that it no
2450		longer needs any knowledge of the compiler.  It exclusively deals
2451		with the details of binfile layout.  It no longer invokes the
2452		compiler (for the purpose of creating new prospective binfile
2453		content), and it no longer has any knowledge of how to interpret
2454		pickles.
2455
2456		Structure Compile (: COMPILE) has been stripped down to the bare
2457		essentials of compilation.  It no longer deals with linking/execution.
2458		The interface has been cleaned up considerably.
2459
2460		Utility routines for dealing with linking and execution have been
2461		moved into their own substructures.
2462
2463		(The ultimate goal of these changes is to provide a light-weight
2464		binfile loader/linker (at least for, e.g., stable libraries) that
2465		does not require CM or the compiler to be present.)
2466
2467		CM documentation has been updated to reflect the changes to library
2468		organization.
2469
2470		----------------------------------------------------------------------
2471		Name: Matthias Blume
2472		Date: 2001/07/10 17:30:00 EDT
2473		Tag: Release_110_34
2474		Description:
2475
2476		Minor tweak to 110.34 (re-tagged):
2477
2478		- README.html file added to CVS repository
2479		- runtime compiles properly under FreeBSD 3.X and 4.X
2480
2481		----------------------------------------------------------------------
2482		Name: Matthias Blume
2483		Date: 2001/07/10 17:30:00 EDT
2484		Tag: Release_110_34
2485		Description:
2486
2487		New version number (110.34). New bootfiles.
2488
2489		----------------------------------------------------------------------
2490		Name: Matthias Blume
2491		Date: 2001/07/09 16:00:00 EDT
2492		Tag: blume-20010709-more-varargs
2493		Description:
2494
2495		I changed the handling of varargs in ml-nlffigen again:
2496		The ellipsis ... will now simply be ignored (with an accompanying warning).
2497
2498		The immediate effect is that you can actually call a varargs function
2499		from ML -- but you can't actually supply any arguments beyond the ones
2500		specified explicitly.  (For example, you can call printf with its format
2501		string, but you cannot pass additional arguments.)
2502
2503		This behavior is only marginally more useful than the one before, but
2504		it has the advantage that a function or, more importantly, a function
2505		type never gets dropped on the floor, thus avoiding follow-up problems with
2506		other types that refer to the offending one.
2507
2508		----------------------------------------------------------------------
2509		Name: Matthias Blume
2510		Date: 2001/07/09 11:25:00 EDT
2511		Tag: blume-20010709-varargs
2512		Description:
2513
2514		1. ckit-lib.cm now exports structure Error
2515		2. ml-nlffigen reports occurences of "..." (i.e., varargs function types)
2516		with a warning accompanied by a source location.  Moreover, it
2517		merely skips the offending function or type and proceeds with the
2518		rest of its work.u  As a result, one can safely feed C code containing
2519		"..." to ml-nlffigen.
2520		3. There are some internal improvements to CM, providing slightly
2521		more general string substitutions in the tools subsystem.
2522
2523		----------------------------------------------------------------------
2524		Name: Matthias Blume
2525		Date: 2001/06/27 15:10:00 EDT
2526		Tag: blume-20010627-concur
2527		Description:
2528
2529		Fixed a small bug in CM's handling of parallel compilation.
2530		(You could observe the bug by Control-C-interrupting an ordinary
2531		CMB.make or CM.stabilize and then attaching some compile servers.
2532		The result was that all of a sudden the previously interrupted
2533		compilation would continue on its own.  This was because of
2534		an over-optimization: CM did not bother to clean out certain queues
2535		when no servers were attached "anyway", resulting in the contents
2536		of these queues to grab control when new servers did get attached.)
2537
2538		There is also another minor update to the CM manual.
2539
2540		----------------------------------------------------------------------
2541		Name: Matthias Blume
2542		Date: 2001/06/26 16:15:00 EDT
2543		Tag: blume-20010626-cmdoc
2544		Description:
2545
2546		Minor typo fixed in CM manual (syntax diagram for libraries).
2547
2548		----------------------------------------------------------------------
2549		Name: Matthias Blume
2550		Date: 2001/06/25 22:55:00 EDT
2551		Tag: blume-20010625-x86pc
2552		Description:
2553
2554		Fixed a nasty bug in the X86 assembly code that caused signal
2555		handlers to fail (crash) randomly.
2556
2557		----------------------------------------------------------------------
2558		Name: Matthias Blume
2559		Date: 2001/06/25 12:05:00 EDT
2560		Tag: blume-20010625-nlffigen
2561		Description:
2562
2563		This update fixes a number of minor bugs in ml-nlffigen as reported by
2564		Nick Carter <nbc@andrew.cmu.edu>.
2565
2566		1. Silly but ok typedefs of the form "typedef void myvoid;" are now accepted.
2567		2. Default names for generated files are now derived from the name of
2568		the C file *without its directory*.  In particular, this causes generated
2569		files to be placed locally even if the C file is in some system directory.
2570		3. Default names for generated signatures and structures are also derived
2571		from the C file name without its directory.  This avoids silly things
2572		like "structure GL/GL".
2573		(Other silly names are still possible because ml-nlffigen does not do
2574		a thorough check of whether generated names are legal ML identifiers.
2575		When in doubt, use command line arguments to force particular names.)
2576
2577		----------------------------------------------------------------------
2578		Name: Matthias Blume
2579		Date: 2001/06/21 12:25:00 EDT
2580		Tag: blume-20010621-eXene
2581		Description:
2582
2583		eXene now compiles and (sort of) works again.
2584
2585		The library name (for version > 110.33) is $/eXene.cm.
2586
2587		I also added an new example in src/eXene/examples/nbody.  See the
2588		README file there for details.
2589
2590		----------------------------------------------------------------------
2591		Name: Matthias Blume
2592		Date: 2001/06/20 16:40:00 EDT
2593		Tag: blume-20010620-cml
2594		Description:
2595
2596		CML now compiles and works again.
2597
2598		Libraries (for version > 110.33):
2599
2600		$cml/cml.cm            Main CML library.
2601		$cml/basis.cm          CML's version of $/basis.cm.
2602		$cml/cml-internal.cm   Internal helper library.
2603		$cml/core-cml.cm       Internal helper library.
2604		$cml-lib/trace-cml.cm  Tracing facility.
2605		$cml-lib/smlnj-lib.cm  CML's version of $/smlnj-lib.cm
2606
2607		The installer (config/install.sh) has been taught how to properly
2608		install this stuff.
2609
2610		----------------------------------------------------------------------
2611		Name: Matthias Blume
2612		Date: 2001/06/19 17:55:00 EDT
2613		Tag: blume-20010619-instantiate
2614		Description:
2615
2616		This un-breaks the fix for bug 1432.
2617		(The bug was originally fixed in 110.9 but I broke it again some
2618		time after that.)
2619
2620		----------------------------------------------------------------------
2621		Name: Matthias Blume
2622		Date: 2001/06/19 17:25:00 EDT
2623		Tag: blume-20010619-signals
2624		Description:
2625
2626		This should (hopefully) fix the long-standing signal handling bug.
2627		(The runtime system was constructing a continuation record with an
2628		incorrect descriptor which would cause the GC to drop data on the floor...)
2629
2630		----------------------------------------------------------------------
2631		Name: Matthias Blume
2632		Date: 2001/06/15 15:05:00 EDT
2633		Tag: blume-20010615-moresparc
2634		Description:
2635
2636		Here is a short late-hour update related to Sparc c-calls:
2637
2638		-- made handling of double-word arguments a bit smarter
2639
2640		-- instruction selection phase tries to collapse certain clumsily
2641		constructed ML-Trees; typical example:
2642
2643		ADD(ty,ADD(_,e,LI d1),LI d2)  ->  ADD(ty,e,LI(d1+d2))
2644
2645		This currently has no further impact on SML/NJ since mlriscGen does
2646		not seem to generate such patterns in the first place, and c-calls
2647		(which did generate them in the beginning) has meanwhile been fixed
2648		so as to avoid them as well.
2649
2650		----------------------------------------------------------------------
2651		Name: Matthias Blume
2652		Date: 2001/06/15 15:05:00 EDT
2653		Tag: blume-20010615-sparc
2654		Description:
2655
2656		The purpose of this update is to provide an implementation of NLFFI
2657		on Sparc machines.
2658
2659		Here are the changes in detail:
2660
2661		* src/MLRISC/sparc/c-calls/sparc-c-calls.sml is a new file containing
2662		the Sparc implementation of the c-calls API.
2663		* The Sparc backend of SML/NJ has been modified to uniformely use %fp
2664		for accessing the ML frame.  Thus, we have a real frame pointer and
2665		can freely modify %sp without need for an omit-frame-ptr phase.
2666		The vfp logic in src/compiler/CodeGen/* has been changed to accomodate
2667		this case.
2668		* ml-nlffigen has been taught to produce code for different architectures
2669		and calling conventions.
2670		* In a way similar to what was done in the x86 case, the Sparc
2671		backend uses its own specific extension to mltree.  (For example,
2672		it needs to be able to generate UNIMP instructions which are part
2673		of the calling convention.)
2674		* ml-nlffi-lib was reorganized to make it more modular (in particular,
2675		to make it easier to plug in new machine- and os-dependent parts).
2676
2677		There are some other fairly unrelated bug fixes and cleanups as well:
2678
2679		* I further hacked the .cm files for MLRISC tools (like MDLGen) so
2680		that they properly share their libraries with existing SML/NJ libraries.
2681		* I fixed a minor cosmetic bug in CM, supressing certain spurious
2682		follow-up error messages.
2683		* Updates to CM/CMB documentation.
2684
2685		TODO items:
2686
2687		* MLRISC should use a different register as its asmTemp on the Sparc.
2688		(The current %o2 is a really bad choice because it is part of the
2689		calling conventions, so things might interfere in unexpected ways.)
2690
2691		----------------------------------------------------------------------
2692		Name: Matthias Blume
2693		Date: 2001/06/07
2694		Tag: blume-20010607-calls
2695		Description:
2696
2697		A number of internal changes related to C calls and calling conventions:
2698
2699		1. ML-Tree CALL statements now carry a "pops" field.  It indicates the
2700		number of bytes popped implicitly (by the callee).  In most cases
2701		this field is 0 but on x86/win32 it is some non-zero value.  This
2702		is information provided for the benefit of the "omit-frameptr" pass.
2703		2. The CALL instruction on the x86 carries a similar "pops" field.
2704		The instruction selection phase copies its value from the ML-Tree
2705		CALL statement.
2706		3. On all other architectures, the instruction selection phase checks
2707		whether "pops=0" and complains if not.
2708		4. The c-calls implementation for x86 now accepts two calling conventions:
2709		"ccall" and "stdcall".  When "ccall" is selected, the caller cleans
2710		up after the call and pops is set to 0.  For "stdcall", the caller
2711		does nothing, leaving the cleanup to the callee; pops is set to
2712		the number of bytes that were pushed onto the stack.
2713		5. The cproto decoder (compiler/Semant/types/cproto.sml) now can
2714		distinguish between "ccall" and "stdcall".
2715		6. The UNIMP instruction has been added to the supported Sparc instruction
2716		set. (This is needed for implementing the official C calling convention
2717		on this architecture.)
2718		7. I fixed some of the .cm files under src/MLRISC/Tools to make them
2719		work with the latest CM.
2720
2721		----------------------------------------------------------------------
2722		Name: Matthias Blume
2723		Date: 2001/06/05 15:10:00 EDT
2724		Tag: blume-20010605-cm-index
2725		Description:
2726
2727		0. The "lambdasplit" parameter for class "sml" in CM has been documented.
2728
2729		1. CM can now generate "index files".  These are human-readable files
2730		that list on a per-.cm-file basis each toplevel symbol defined or
2731		imported.  The location of the index file for
2732		<p>/<d>.cm is <p>/CM/INDEX/<d>.cm.
2733		To enable index-file generation, set CM.Control.generate_index to true
2734		or export an environment-symbol: export CM_GENERATE_INDEX=true.
2735
2736		The CM manual has been updated accordingly.
2737
2738		2. I made some slight modifications to the c-calls API in MLRISC.
2739
2740		a) There is now a callback to support saving/restoring of
2741		dedicated but caller-save registers around the actual call
2742		instruction.
2743		b) One can optionally specify a comment-annotation for the
2744		call instruction.
2745
2746		3. SML/NJ (mlriscGen.sml) uses this new API for the rawccall primop.
2747		(For example, the comment annotation shows the C prototype of
2748		the function being called.)
2749
2750		----------------------------------------------------------------------
2751		Name: Matthias Blume
2752		Date: 2001/06/01 13:30:00 EDT
2753		Tag: blume-20010601-nlffi-cleanup
2754		Description:
2755
2756		This is mostly a cleanup of MLFFI stuff:
2757
2758		- some signature files have been put into a more exposed place
2759		- the ugly 'f type parameter is gone (simplifies types tremendously!)
2760		- ml-nlffigen changed accordingly
2761		- tutorial updated
2762
2763		Other changes:
2764
2765		- author's affiliation in CM manual(s) updated
2766		- some more recognized keywords added to Allen's sml.sty
2767
2768		----------------------------------------------------------------------
2769		Name: Matthias Blume
2770		Date: 2001/05/25 15:30:00 EDT
2771		Tag: blume-20010525-iptr
2772		Description:
2773
2774		- put the official 110.33-README (as it appears on the ftp server) under
2775		CVS
2776		- fixed a small bug related to incomplete pointer types in
2777		ml-nlffigen
2778		- small cosmetic change to the ml-nlffi-lib's "arr" type constructor
2779		(it does not need the 'f type parameter)
2780
2781		----------------------------------------------------------------------
2782		Name: Matthias Blume
2783		Date: 2001/05/23 14:30:00 EDT
2784		Tag: Release_110_33
2785		Description:
2786
2787		New version number (110.33).  New bootfiles.
2788
2789		----------------------------------------------------------------------
2790		Name: Matthias Blume
2791		Date: 2001/05/22 18:06:00 EDT
2792		Tag: blume-20010522-targets
2793		Description:
2794
2795		Made install.sh use file config/targets.customized if it exists, falling
2796		back to config/targets if it doesn't.  This way one can have a customized
2797		version of the targets file without touching the "real thing", thus
2798		eliminating the constant fear of accidentally checking something bogus
2799		back into the CVS repository...  (File config/targets.customized must
2800		not be added to the repository!)
2801
2802		----------------------------------------------------------------------
2803		Name: Matthias Blume
2804		Date: 2001/05/22 16:30:00 EDT
2805		Tag: blume-20010522-minitut
2806		Description:
2807
2808		1. Bug fix in ml-nlffigen; now (hopefully) correctly handling
2809		struct returns.
2810		2. Added src/ml-nlffi-lib/Doc/mini-tutorial.txt.  This is some very
2811		incomplete, preliminary documentation for NLFFI.
2812
2813		----------------------------------------------------------------------
2814		Name: Matthias Blume
2815		Date: 2001/05/14 11:30:00 EDT
2816		Tag: blume-20010514-script
2817		Description:
2818
2819		Some bugs in install script fixed.
2820
2821		In addition to that I also made a slight change to the NLFFI API:
2822		Functors generated by ml-nlffigen now take the dynamic library as a
2823		straight functor argument, not as a suspended one.  (The original
2824		functor code used to force the suspension right away anyway, so there
2825		was nothing gained by this complication of the interface.)
2826
2827		----------------------------------------------------------------------
2828		Name: Matthias Blume
2829		Date: 2001/05/11 14:35:00 EDT
2830		Tag: blume-20010511-ml-nlffi
2831		Description:
2832
2833		I finally took the plunge and added my new FFI code to the main
2834		repository.  For x86-linux it is now ready for prime-time.
2835
2836		There are two new subdirectories of "src":
2837
2838		- ml-nlffi-lib:
2839		The utility library for programs using the FFI interface.
2840		Here is the implementation of $/c.cm and its associated low-level
2841		partners $/c-int.cm and $/memory.cm.
2842		- ml-nlffigen:
2843		A stand-alone program for generating ML glue code from C source
2844		code.
2845
2846		Building ml-nlffigen requires $/ckit-lib.cm.
2847
2848		The config/install.sh script has been updates to do the Right Thing
2849		(hopefully).
2850
2851		Notice that the source tree for the C-Kit will not be put under "src"
2852		but directly under the installation root directory.  (This is the
2853		structure that currently exists on the CVS server when you check out
2854		module "sml".)  Fortunately, config/install.sh knows about this oddity.
2855
2856		Bugs: No documentation yet.
2857
2858		----------------------------------------------------------------------
2859		Name: Matthias Blume
2860		Date: 2001/05/09 16:35:00 EDT
2861		Tag: blume-20010509-cpscontract
2862		Description:
2863
2864		Fixed a bug in the accounting code in cpsopt/contract.sml.  (The
2865		wrapper/unwrapper elimination did not decrement usage counts and some
2866		dead variables got overlooked by the dead-up logic.)
2867
2868		----------------------------------------------------------------------
2869		Name: Lal George
2870		Date: 2001/05/08  17:26:09 EDT
2871		Tag: george-20010508-omit-frameptr
2872		Description:
2873
2874		Changes to implement the omit-frame-pointer optimization to support
2875		raw C calls. For now, there is only support on the Intel x86, but
2876		other architectures will follow as more experience is gained with this.
2877
2878
2879		----------------------------------------------------------------------
2880		Name: Matthias Blume
2881		Date: 2001/05/07 14:40:00 EDT
2882		Tag: blume-20010507-proxies
2883		Description:
2884
2885		I made into "proxy libraries" all libraries that qualify for such a
2886		change.  (A qualifying library is a library that has another library or
2887		groups as its sole member and repeats that member's export list
2888		verbatim.  A proxy library avoids this repetition by omitting its export
2889		list, effectively inheriting the list that its (only) member exports.
2890		See the CM manual for more explanation.)
2891		The main effect is that explicit export lists for these libraries
2892		do not have to be kepts in sync, making maintenance a bit easier.
2893
2894		I also added copyright notices to many .cm-files.
2895
2896		Last but not least, I made a new set of bootfiles.
2897
2898		----------------------------------------------------------------------
2899		Name: Matthias Blume
2900		Date: 2001/05/04 17:00:00 EDT
2901		Tag: blume-20010504-cm-lsplit
2902		Description:
2903
2904		0. John merged pending changes to $/smlnj-lib.cm
2905
2906		1. Allen's previous change accidentally backed out of one of Lal's
2907		earlier changes.  I undid this mistake (re-introducing Lal's change).
2908
2909		2. I used the new topOrder' function from graph-scc.sml (from $/smlnj-lib.cm)
2910		within the compiler where applicable.  There is some code simplification
2911		because of that.
2912
2913		3. The "split" phase (in FLINT) is now part of the default list of phases.
2914		Compiler.Control.LambdaSplitting.* can be used to globally control the
2915		lambda-splitting (cross-module-inlining) engine.  In addition to that,
2916		it can now also be controlled on a per-source basis: CM has been taught
2917		a new tool parameter applicable to ML source files.
2918
2919		- To turn lambda-splitting off completely:
2920		local open Compiler.Control.LambdaSplitting in
2921		val _ = set Off
2922		end
2923		- To make "no lambda-splitting" the global default (but allow per-source
2924		overriding); this is the initial setting:
2925		local open Compiler.Control.LambdaSplitting in
2926		val _ = set (Default NONE)
2927		end
2928		- To make "lambda-splitting with aggressiveness a" the global default
2929		(and allow per-source overriding):
2930		local open Compiler.Control.LambdaSplitting in
2931		val _ = set (Default (SOME a))
2932		end
2933
2934		- To turn lambda-splitting off for a given ML souce file (say: a.sml)
2935		write (in the respective .cm-file):
2936		a.sml (lambdasplitting:off)
2937		- To turn lambda-splitting for a.sml on with minimal aggressiveness:
2938		a.sml (lambdasplitting:on)
2939		- To turn lambda-splitting for a.sml on with aggressiveness <a> (where
2940		<a> is a decimal non-negative integer):
2941		a.sml (lambdasplitting:<a>)
2942		- To turn lambda-splitting for a.sml on with maximal aggressiveness:
2943		a.sml (lambdasplitting:infinity)
2944		- To use the global default for a.sml:
2945		a.sml (lambdasplitting:default)
2946		or simply
2947		a.sml
2948
2949		----------------------------------------------------------------------
2950		Name: Allen Leung
2951		Date: 2001/05/04 01:57:00 EDT
2952		Tag: leunga-20010504-sync
2953		Description:
2954
2955		MLRISC features.
2956
2957		1. Fix to CMPXCHG instructions.
2958		2. Changed RA interface to allow annotations in callbacks.
2959		3. Added a new method to the stream interface to allow annotations updates.
2960
2961		----------------------------------------------------------------------
2962		Name: Matthias Blume
2963		Date: 2001/05/01 11:45:00 EDT
2964		Tag: blume-20010501-pcedittmp
2965		Description:
2966
2967		Changed install.sh to use the current working directory instead of
2968		/usr/tmp for a temporary file (pcedittmp).  The previous choice
2969		of /usr/tmp caused trouble with MacOS X because of file premission
2970		problems.
2971
2972		----------------------------------------------------------------------
2973		Name: Matthias Blume
2974		Date: 2001/04/20 11:10:00 EDT
2975		Tag: blume-20010420-inMLflag
2976		Description:
2977
2978		- added vp_limitPtrMask to vproc-state.h
2979		(for use by the raw-C-calls mechanism to implement proper interrupt
2980		handling)
2981		- made the ML compiler aware of various data-structure offsets so it
2982		can generate code for accessing the vp_inML flag and vp_limitPtrMask
2983		- tweaked mlriscGen.sml to have it emit interrupt-handling code for
2984		raw C-calls
2985
2986		----------------------------------------------------------------------
2987		Name: Lal George
2988		Date: 2001/04/20 09:15:28 EDT
2989		Tag: george-20010420-macosX
2990		Description:
2991
2992		- Changes to port to Mac OS X; Darwin.
2993
2994		- In the process I found that sqrt was broken on the PPC, because the
2995		fsqrt instruction is not implemented.
2996
2997		----------------------------------------------------------------------
2998		Name: Matthias Blume
2999		Date: 2001/04/18 12:45:00 EDT
3000		Tag: blume-20010418-ccalls
3001		Description:
3002
3003		- fixed two off-by-4 errors in the x86-specific c-calls implementation
3004		(this bug prevented structure arguments containing pointers from being
3005		passed correctly)
3006		- changed the raw-C-call code in mlriscGen.sml in such a way that
3007		structure arguments are represented as a pointer to the beginning
3008		of the structure (instead of having a series of synthesized arguments,
3009		one for each structure member)
3010
3011		- made makeml script's verbosity level configurable via environment
3012		variable (MAKEML_VERBOSITY)
3013
3014		- eliminated placeholder implementations for f32l, w16s, i16s, and f32s
3015		in rawmem-x86.sml; we are now using the real thing
3016
3017		----------------------------------------------------------------------
3018		Name: Matthias Blume
3019		Date: 2001/03/22 16:25:00 EST
3020		Tag: blume-20010322-bootfiles
3021		Description:
3022
3023		Created a new set of bootfiles (for your automatic installation convenience).
3024
3025		----------------------------------------------------------------------
3026		Name: Matthias Blume
3027		Date: 2001/03/22 15:10:00 EST
3028		Tag: blume-20010322-rawmem-parcm
3029		Description:
3030
3031		1. All "raw memory access" primitives for the new FFI are implemented now
3032		(at least on the x86).
3033		2. Some further cleanup of CM's parallel make mechanism.
3034
3035		----------------------------------------------------------------------
3036		Name: Matthias Blume
3037		Date: 2001/03/19 17:53:00 EST
3038		Tag: blume-20010319-parallel
3039		Description:
3040
3041		Parallel make (using compile servers) now works again.
3042
3043		To this end, CM.stabilize and CMB.make have been modified to work in
3044		two passes when compile servers are attached:
3045		1. Compile everything, do not perform stabilization; this pass
3046		uses compile servers
3047		2. Stabilize everything; this pass does not use compile servers
3048		If there are no compile servers, the two passes are combined into one
3049		(as before).  Splitting the passes increases the inherent parallelism
3050		in the dependency graph because the entire graph including all
3051		libraries is available at the same time.  This, in turn, improves
3052		server utilization.  The downside is that the master process will
3053		have to do some extra work after compilation is done (because for
3054		technical reasons it must re-read all the binfiles during stabilization).
3055
3056		----------------------------------------------------------------------
3057		Name: Matthias Blume
3058		Date: 2001/03/16 12:22:00 EST
3059		Tag: blume-20010316-bootfiles
3060		Description:
3061
3062		Created a new set of bootfiles (for your automatic installation convenience).
3063
3064		----------------------------------------------------------------------
3065		Name: Matthias Blume
3066		Date: 2001/03/16 11:00:00 EST
3067		Tag: blume-20010316-MLTREE-fixup
3068		Description:
3069
3070		This is a minor fixup for an (untagged) earlier commit by Allen.
3071		(A file was missing).
3072
3073		----------------------------------------------------------------------
3074		Name: Allen Leung
3075		Date: Mon Mar  5 18:54:57 EST 2001
3076		Tag: leunga-20010305-cut-support
3077
3078		1. New support for alternative control-flow in MLTREE.
3079		Currently we support
3080
3081		FLOW_TO(CALL ...., [k1,...,kn])
3082
3083		This is needed for 'cuts to' in C-- and try/handle-like constructs
3084		in Moby
3085
3086		New assembler flag "asm-show-cutsto" to turn on control-flow debugging.
3087
3088		2. Register Allocator
3089
3090		Changes in interface [from Fermin, John]
3091
3092		3. Alpha 8-bit SLL support [Fermin]
3093
3094		4. All architectures
3095
3096		A new module (ClusterExpandCopies) for expanding parallel copies.
3097
3098		----------------------------------------------------------------------
3099		Name: Allen Leung
3100		Date: 2001/02/27 23:07:00 EST
3101		Tag: leunga-20010227-minor-stuff
3102
3103		1. Alpha bug fix for CMOVNE
3104		2. Handle mltree COND(..,FCMP ...,...)
3105		3. Bug fix in simplifier
3106
3107		----------------------------------------------------------------------
3108		Name: Matthias Blume
3109		Date: 2001/01/30 17:50:00 EST
3110		Tag: blume-20010130-sync
3111		Description:
3112
3113		This is just a minor update to sync my devel branch with the main brach.
3114		The only visible change is the addition of some README files.
3115
3116		----------------------------------------------------------------------
3117		Name: Matthias Blume
3118		Date: 2001/01/12 23:30:00 JST
3119		Tag: blume-20010112-bootfiles
3120		Description:
3121
3122		Made a new set of bootfiles that goes with the current state of the
3123		repository.
3124
3125		----------------------------------------------------------------------
3126		Name: Matthias Blume
3127		Date: 2001/01/12 21:20:00 JST
3128		Tag: blume-20010112-sync
3129		Description:
3130
3131		I am just flushing out some minor changes that had accumulated in
3132		my private branch in order to sync with the main tree.  (This is
3133		mainly because I had CVS trouble when trying to merge _into_ my
3134		private branch.)
3135
3136		Most people should be completely unaffected by this.
3137
3138		----------------------------------------------------------------------
3139		Name: Allen Leung
3140		Date: Thu Jan 11 21:03:00 EST 2001
3141		Tag: leunga-20010111-labexp=mltree
3142		Description:
3143
3144		1.  Removed the type LabelExp and replace it by MLTree.
3145		2.  Rewritten mltree-simplify with the pattern matcher tool.
3146		3.  There were some bugs in alpha code generator which would break
3147		64-bit code generation.
3148		4.  Redo the tools to generate code with the
3149		5.  The CM files in MLRISC (and in src/system/smlnj/MLRISC)
3150		are now generated by perl scripts.
3151
3152		----------------------------------------------------------------------
3153		Name: Matthias Blume
3154		Date: 2001/01/10 21:55:00 JST
3155		Tag: blume-20010110-rcc
3156		Description:
3157
3158		The RCC stuff now seems to work (but only on the x86).
3159		This required hacking of the c-calls interface (and -implementation) in
3160		MLRISC.
3161
3162		Normal compiler users should be unaffected.
3163
3164		----------------------------------------------------------------------
3165		Name: Matthias Blume
3166		Date: 2001/01/09 01:20:00 JST
3167		Tag: blume-20010109-rcc
3168		Description:
3169
3170		This is a fairly big patch, flushing out a large number of pending
3171		changes that I made to my development copy over the last couple of days.
3172
3173		Of practical relevance at this moment is a workaround for a pickling
3174		bug that Allen ran into the other day.  The cause of the bug itself is
3175		still unknown and it might be hard to fix it properly, but the
3176		workaround has some merits of its own (namely somewhat reducing pickling
3177		overhead for certain libraries).  Therefore, I think this solution should
3178		be satisfactory at this time.
3179
3180		The rest of the changes (i.e., the vast majority) has to do with my
3181		ongoing efforts of providing direct support for C function calls from
3182		ML.  At the moment there is a new primop "RAW_CCALL", typing magic
3183		in types/cproto.sml (invoked from FLINT/trans/translate.sml), a new
3184		case in the FLINT CPS datatype (RCC), changes to cps/convert.sml to
3185		translate uses of RAW_CCALL into RCC, and changes to mlriscGen.sml to
3186		handle RCC.
3187
3188		The last part (the changes to mlriscGen.sml) are still known to be
3189		wrong on the x86 and not implemented on all other architectures.  But
3190		the infrastructure is in place. I had to change a few functor
3191		signatures in the backend to be able to route the CCalls interface
3192		from MLRISC there, and I had to specialize the mltree type (on the
3193		x86) to include the necessary extensions. (The extensions themselves
3194		were already there and redy to go in MLRISC/x86).
3195
3196		Everything should be very happy as soon as someone helps me with
3197		mlriscGen.sml...
3198
3199		In any case, nothing of this should matter to anyone as long as the
3200		new primop is not being used (which is going to be the case unless you
3201		find it where I hid it :). The rest of the compiler is completely
3202		unaffected.
3203
3204		----------------------------------------------------------------------
3205		Name: Matthias Blume
3206		Date: 2001/01/05 00:30:00 JST
3207		Tag: blume-20010105-primops
3208		Description:
3209
3210		Added some experimental support for work that I am doing right now.
3211		These changes mostly concern added primops, but there is also a new
3212		experimental C library in the runtime system (but currently not enabled
3213		anywhere except on Linux/X86).
3214
3215		In the course of adding primops (and playing with them), I discovered that
3216		Zhong's INL_PRIM hack (no type info for certain primops) was, in fact, badly
3217		broken.  (Zhong was very right he labeled this stuff as "major gross hack".)
3218		To recover, I made type information in INL_PRIM mandatory and changed
3219		prim.sml as well as built-in.sml accordingly.  The InLine structure now
3220		has complete, correct type information (i.e., no bottom types).
3221
3222		Since all these changes mean that we need new binfiles, I also bumped the
3223		version number to 110.32.1.
3224
3225		----------------------------------------------------------------------
3226		Name: Matthias Blume
3227		Date: 2000/12/30 22:10:00 JST
3228		Tag: blume-20001230-various
3229		Description:
3230
3231		Added proxy libraries for MLRISC and let MLRISC libraries refer
3232		to each other using path anchors.  (See CM manual for explanation.)
3233
3234		Updated CM documentation.
3235
3236		Fixed some bugs in CM.
3237
3238		Implemented "proxy" libraries (= syntactic sugar for CM).
3239
3240		Added "-quiet" option to makeml and changed runtime system accordingly.
3241
3242		Added cleanup handler for exportML to reset timers and compiler stats.
3243
3244		----------------------------------------------------------------------
3245		Name: Lal George
3246		Date: 2000/12/22 22:22:58 EST 2000
3247		Tag: Release_110_32
3248		Description:
3249
3250		Infinite precision used throughout MLRISC.
3251		see MLRISC/mltree/machine-int.sig
3252
3253		----------------------------------------------------------------------
3254		Name: Matthias Blume
3255		Date: 2000/12/22 23:16:00 JST
3256		Tag: blume-20001222-warn
3257		Description:
3258
3259		Corrected wording and formatting of some CM warning message which I
3260		broke in my previous patch.
3261
3262		----------------------------------------------------------------------
3263		Name: Matthias Blume
3264		Date: 2000/12/22 21:20:00 JST
3265		Tag: blume-20001222-anchorenv
3266		Description:
3267
3268		Fixed CM's handling of anchor environments in connection with CMB.make.
3269
3270		----------------------------------------------------------------------
3271		Name: Matthias Blume
3272		Date: 2000/12/22 13:15:00 JST
3273		Tag: blume-20001222-cleanup
3274		Description:
3275
3276		Removed src/cm/ffi which does not (and did not) belong here.
3277
3278		----------------------------------------------------------------------
3279		Name: Matthias Blume
3280		Date: 2000/12/21 23:55:00 JST
3281		Tag: blume-20001221-exn
3282		Description:
3283
3284		Probably most important: CM no longer silently swallows all exceptions
3285		in the compiler.
3286		Plus: some other minor CM changes.  For example, CM now reports some
3287		sizes for generated binfiles (code, data, envpickle, lambdapickle).
3288
3289		----------------------------------------------------------------------
3290		Name: Matthias Blume
3291		Date: 2000/12/15 00:01:05 JST
3292		Tag: blume-20001215-dirtool
3293		Description:
3294
3295		- "dir" tool added.
3296		- improvements and cleanup to Tools structure
3297		- documentation updates
3298
3299		----------------------------------------------------------------------
3300		Name: Allen Leung
3301		Date: Thu Dec 14 03:45:24 EST 2000
3302		Description:
3303		Tag:  leunga-20001214-int-inf
3304		Description:
3305
3306		In IntInf, added these standard functions, which are missing from our
3307		implementation:
3308
3309		andb : int * int -> int
3310		xorb : int * int -> int
3311		orb  : int * int -> int
3312		notb : int -> int
3313		<<   : int * word -> int
3314		~>>  : int * word -> int
3315
3316		Not tested, I hope they are correct.
3317
3318		----------------------------------------------------------------------
3319		Name: Allen Leung
3320		Date: Fri Dec  8 19:23:26 EST 2000
3321		Description:
3322		Tag:  leunga-20001208-nowhere
3323		Description:
3324
3325		Slight improvements to the 'nowhere' tool to handle OR-patterns,
3326		to generate better error messages etc.  Plus a brief manual.
3327
3328		----------------------------------------------------------------------
3329		Name: Lal George
3330		Date: 2000/12/08 09:54:02 EST 2000
3331		Tag: Release_110_31
3332		Description:
3333
3334		- Version 110.31
3335		----------------------------------------------------------------------
3336		Name: Allen Leung
3337		Date: Thu Dec  7 22:01:04 EST 2000
3338		Tag:  leunga-20001207-cell-monster-hack
3339		Description:
3340
3341		Major MLRISC internal changes.  Affect all clients.
3342		Summary:
3343
3344		1.  Type CELLS.cell = int is now replaced by a datatype.
3345		As a result, the old regmap is now gone.  Almost all interfaces
3346		in MLRISC change as a consequence.
3347
3348		2.  A new brand version of machine description tool (v3.0) that generates
3349		modules expecting the new interface.  The old version is removed.
3350
3351		3.  The RA interface has been further abstracted into two new functors.
3352		RISC_RA and X86RA.  These functors have much simpler interfaces.
3353		[See also directory MLRISC/demo.]
3354
3355		4.  Some other new source->source code generation tools are available:
3356
3357		a. MLRISC/Tools/RewriteGen -- generate rewriters from rules.
3358		b. MLRISC/Tools/WhereGen -- expands conditional pattern matching rules.
3359		I use this tool to generate the peephole optimizers---with the new
3360		cell type changes, peephole rules are becoming difficult to write
3361		without conditional pattern matching.
3362
3363		5.  More Intmap -> IntHashTable change.  Previous changes by Matthias didn't
3364		cover the entire MLRISC source tree so many things broke.
3365
3366		6.  CM files have been moved to the subdirectory MLRISC/cm.
3367		They are moved because there are a lot of them and they clutter up the
3368		root dir.
3369
3370		7.  More detailed documentation to come...
3371
3372		NOTE: To rebuild from 110.30 (ftp distribution), you'll have to do
3373		a makeml -rebuild first.  This is because of other other
3374		changes that Matthias has made (see below).
3375
3376
3377		----------------------------------------------------------------------
3378		Name: Matthias Blume
3379		Date: 2000/11/30 23:12:00 JST
3380		Tag: blume-20001130-filereorg
3381		Description:
3382
3383		Some manual updates and some file reorganizations in CM.
3384
3385		----------------------------------------------------------------------
3386		Name: Matthias Blume
3387		Date: 2000/11/24 17:45:00 JST
3388		Tag: blume-20001124-link
3389		Description:
3390
3391		Drastically improved link traversal code for the case that the dynamic
3392		value was already loaded at bootstrap time.  As a result, CM and CMB
3393		now both load blazingly fast -- even on a very slow machine.  Also,
3394		memory consumption has been further reduced by this.
3395
3396		Warning: The format of the PIDMAP file has changed.  THerefore, to
3397		bootstrap you have to do this:
3398
3399		1. Run CMB.make
3400		2. Make a symbolic link for the boot directory:
3401		ln -s sml.boot.ARCH-OS xxx
3402		3. "Rebuild" the boot directory:
3403		./makeml -boot xxx -rebuild sml ; rm xxx
3404		4. Boot normally:
3405		./makeml
3406
3407		----------------------------------------------------------------------
3408		Name: Matthias Blume
3409		Date: 2000/11/21 21:20:00 JST
3410		Tag: blume-20001121-tools
3411		Description:
3412
3413		Continued hacking on autoloading problem -- with success this time.
3414		Also changed tool-plugin mechanism.  See new CM manual.
3415
3416		----------------------------------------------------------------------
3417		Name: Matthias Blume
3418		Date: 2000/11/19 14:30:00 JST
3419		Tag:  blume-20001119-autoload
3420		Description:
3421
3422		Some hacking to make autoloading faster.  Success for CMB, no success
3423		so far for CM.  There is a reduced structure CM' that autoloads faster.
3424		(This is a temporary, non-documented hack to be eliminated again when
3425		the general problem is solved.)
3426
3427		----------------------------------------------------------------------
3428		Name: Matthias Blume
3429		Date: 2000/11/17 14:10:00 JST
3430		Tag: blume-20001117-pickle-lib
3431		Description:
3432
3433		1. Eliminated comp-lib.cm
3434		2. Made pickle-lib.cm
3435		3. Eliminated all uses of intset.sml (from comp-lib.cm)
3436		4. Replaced all uses of intmap.{sig,sml} (from comp-lib.cm) with
3437		equivalent constructs from smlnj-lib.cm (INtHashTable).
3438		5. Point 4. also goes for those uses of intmap.* in MLRISC.
3439		Duplicated intmap modules thrown out.
3440		6. Hunted down all duplicated SCC code and replaced it with
3441		equivalent stuff (GraphSCCFn from smlnj-lib.cm).
3442		7. Rewrote Feedback module.
3443		8. Moved sortedlist.sml into viscomp-lib.cm.  Eventually it
3444		should be thrown out and equivalent modules from smlnj-lib.cm
3445		should be used (IntRedBlackSet, IntListSet, ...).
3446
3447		Confirmed that compiler compiles to fixpoint.
3448
3449		----------------------------------------------------------------------
3450		Name: Allen Leung
3451		Date: 2000/11/10 18:00:00
3452		Tag: leunga-20001110-new-x86-fp
3453
3454		A new x86 floating point code generator has been added.
3455		By default this is turned off.  To turn this on, do:
3456
3457		CM.autoload "$smlnj/compiler.cm";
3458		Compiler.Control.MLRISC.getFlag "x86-fast-fp" := true;
3459
3460		Changes:
3461
3462		1.  Changed FTAN to FPTAN so that the assembly output is correct.
3463		2.  Changed the extension callback for FTANGENT to generate:
3464
3465		fptan
3466		fstp  %st(0)
3467		instead of
3468		fptan
3469		fstpl ftempmem
3470
3471		3.  Numerous assembly fixes for x86.
3472
3473		5.  Cleaned up the machine code output module x86/x86MC.sml and added
3474		support for a whole bunch of instructions and addressing modes:
3475
3476		fadd/fsub/fsubr/fmul/fdiv/fdivr  %st, %st(n)
3477		faddp/fsubp/fsubrp/fmulp/fdivp/fdivrp  %st, %st(n)
3478		fadd/fsub/fsubr/fmul/fdiv/fdivr  %st(n), %st
3479		fiadd/fisub/fisubr/fimul/fidiv/fidivr mem
3480		fxch %st(n)
3481		fld %st(n)
3482		fst %st(n)
3483		fst mem
3484		fstp %st(n)
3485		fucom %st(n)
3486		fucomp %st(n)
3487
3488		All these are now generated when the fast fp mode is turned on.
3489
3490		6.  Removed the dedicated registers %st(0), ..., %st(7) from X86CpsRegs
3491
3492		----------------------------------------------------------------------
3493		Name: Matthias Blume
3494		Date: 2000/11/09 11:20:00 JST
3495		Tag: blume-20001109-scc
3496		Description:
3497
3498		Eliminated some code duplication:
3499
3500		1. Added "where" clause to GraphSCCFn in SML/NJ Library.
3501		(Otherwise the functor is useless.)
3502		2. Used GraphSCCFn where SCCUtilFun was used previously.
3503		3. Got rid of SCCUtilFun (in comp-lib.cm).
3504
3505		----------------------------------------------------------------------
3506		Name: Lal George
3507		Date: 2000/11/06 09:02:21 EST 2000
3508		Tag: Release_110_30
3509		Description:
3510
3511		- Version 110.30
3512		----------------------------------------------------------------------
3513		Name: Matthias Blume
3514		Date: 2000/11/04 14:45:00
3515		Tag: blume-20001104-mlbuild
3516		Description:
3517
3518		- Made ml-build faster on startup.
3519		- Documentation fixes.
3520
3521		----------------------------------------------------------------------
3522		Name: Matthias Blume
3523		Date: 2000/11/02 17:00:00 JST
3524		Tag: blume-20001102-condcomp
3525		Description:
3526
3527		- Small tweaks to pickler -- new BOOTFILES!
3528		- Version bumped to 110.29.2.
3529		- Added conditional compilation facility to init.cmi (see comment there).
3530		----------------------------------------------------------------------
3531		Name: Allen Leung
3532		Date: 2000/10/23 19:31:00
3533		Tag: leunga-20001023-demo-ra
3534
3535		1. Minor RA changes that improves spilling on x86 (affects Moby and C-- only)
3536		2. Test programs for the graph library updated
3537		3. Some new MLRISC demo programs added
3538
3539		----------------------------------------------------------------------
3540		Name: Matthias Blume
3541		Date: 2000/08/31 22:15:00 JST
3542		Tag: blume-20001017-errmsg
3543		Description:
3544
3545		More error message grief: Where there used to be no messages, there
3546		now were some that had bogus error regions.  Fixed.
3547
3548		----------------------------------------------------------------------
3549		Name: Matthias Blume
3550		Date: 2000/08/31 17:30:00 JST
3551		Tag: blume-20001017-v110p29p1
3552		Description:
3553
3554		I made a version 110.29.1 with new bootfiles.
3555
3556		Changes:  Modified pickler/unpickler for faster and leaner unpickling.
3557		CM documentation changes and a small bugfix in CM's error reporting.
3558
3559		----------------------------------------------------------------------
3560		Name: Lal George
3561		Date: 2000/09/27 14:42:35 EDT
3562		Tag: george-20000927-nodestatus
3563		Description:
3564
3565		Changed the type of the nodestatus, so that:
3566
3567		SPILLED(~1)             is now SPILLED
3568		SPILLED(m) where m>=0   is now MEMREG(m)
3569		SPILLED(s) where s<~1   is now SPILL_LOC(~s)
3570
3571		----------------------------------------------------------------------
3572		Name: Matthias Blume
3573		Date: 2000/09/07 14:45:00 JST
3574		Tag: blume-20000907-cmerrmsg
3575		Description:
3576
3577		Small tweak to CM to avoid getting ML syntax error messages twice.
3578
3579		----------------------------------------------------------------------
3580		Name: Matthias Blume
3581		Date: 2000/08/31 18:00:00 JST
3582		Tag: blume-20000831-cvsbootfiles
3583		Description:
3584
3585		New URL for boot files (because the 110.29 files on the BL server do
3586		now work correctly with my updated install scripts for yacc and lex).
3587
3588		----------------------------------------------------------------------
3589		Name: Matthias Blume
3590		Date: 2000/08/08 12:33:00 JST
3591		Tag: blume-20000808-manual
3592		Description:
3593
3594		Tiny update to CM manual.
3595
3596		----------------------------------------------------------------------
3597		Name: Allen Leung
3598		Date: 2000/08/7 19:31:00
3599		Tag: leunga-20000807-a-whole-bunch-of-stuff
3600
3601		Moby, C--, SSA, x86, machine descriptions etc.  Should only affect C--
3602		and Mobdy.
3603
3604		1.  x86
3605
3606		a.  Fixes to peephole module by John and Dan.
3607		b.  Assembly fix to SETcc by Allen.
3608		c.  Fix to c-call by John.
3609		d.  Fix to spilling by John.  (This one deals with the missing FSTPT case)
3610		e.  Instruction selection optimization to SETcc as suggested by John.
3611
3612		For example,
3613
3614		MV(32, x, COND(32, CMP(32, LT, a, b), LI 1, LI 0))
3615
3616		should generate:
3617
3618		MOVL a, x
3619		SUBL b, x
3620		SHRL 31, x
3621